├── README.md
├── adversarial-examples
    ├── random_perturbations
    │   ├── README.md
    │   ├── explore_space.py
    │   └── models.py
    └── spatially_transformed
    │   ├── README.md
    │   ├── model_mnist.py
    │   └── stadv.py
├── code
    ├── active.py
    ├── image_list.py
    ├── main.py
    ├── network.py
    ├── prepocess.py
    └── utils.py
└── data
    ├── README.md
    ├── office-home
        ├── Art.txt
        ├── Clipart.txt
        ├── Product.txt
        └── Real_World.txt
    ├── office
        ├── amazon.txt
        ├── dslr.txt
        └── webcam.txt
    └── visda2017
        ├── train_list.txt
        └── validation_list.txt


/README.md:
--------------------------------------------------------------------------------
 1 | # Transferable-Query-Selection
 2 | Code Release for "Transferable Query Selection for Active Domain Adaptation"(CVPR2021)  
 3 | 
 4 | Waiting for code update and document.
 5 | 
 6 | The adversarial-examples refs to https://github.com/sarathknv/adversarial-examples-pytorch
 7 | 
 8 | * **Dataset Download** 
 9 | 
10 | Dataset download：<br />
11 | Office-31: http://people.eecs.berkeley.edu/~jhoffman/domainadapt/ <br />
12 | Office-Home: http://hemanthdv.org/OfficeHome-Dataset/ <br />
13 | VisDA: https://github.com/VisionLearningGroup/taskcv-2017-public/tree/master/classification <br />
14 | Download the dataset according to the instructions on the above, and update the path in each file in the 'data' folder
15 | 
16 | * **Specification of dependencies** 
17 | 
18 | We use the following libraries: pytorch 1.7, torchvision 0.6, numpy 1.18 and matplotlib 3.2. <br />
19 | Pre-trained models resnet-50 can be automatically downloaded from the pytorch community. <br />
20 | 
21 | * **Command** 
22 | 
23 | For Office-31 command: <br />
24 | python3 main.py --gpu 0 --lr 0.1 --batch-size 32 --epochs 50 --source data/office/amazon.txt --source-val data/office/amazon.txt --target data/office/dslr.txt --target-val data/office/dslr.txt --class-num 31 | tee "A_D.log"
25 | 
26 | For Office-Home command: <br />
27 | python3 main.py --gpu 0 --lr 0.1 --epochs 40 --batch-size 32 --source data/office-home/Art.txt --target data/office-home/Clipart.txt --target-val data/office-home/Clipart.txt --class-num 65 | tee "A_C.log"
28 | 
29 | For VisDA command: <br />
30 | python3 main.py --gpu 0 --lr 0.1 --batch-size 32 --epochs 20 --source data/visda2017/train_list.txt --target data/visda2017/validation_list.txt --target-val data/visda2017/validation_list.txt --class-num 12 | tee "vis.log"
31 | 
32 | 


--------------------------------------------------------------------------------
/adversarial-examples/random_perturbations/README.md:
--------------------------------------------------------------------------------
 1 | # Random Perturbations
 2 | 
 3 | 
 4 | From one of the first papers on Adversarial examples - [Explaining and Harnessing Adversarial Examples](https://arxiv.org/abs/1412.6572),
 5 | > The direction of perturbation, rather than the specific point in space, matters most. Space is
 6 | not full of pockets of adversarial examples that finely tile the reals like the rational numbers.  
 7 | 
 8 | This project examines this idea by testing the robustness of a DNN to randomly generated perturbations.
 9 | 
10 | 
11 | 
12 | ## Usage
13 | ```bash
14 | $ python3 explore_space.py --img images/horse.png
15 | ```
16 | 
17 | 
18 | 
19 | ## Demo
20 |  ![fgsm.gif](images/horse_explore_demo.gif)  
21 | 
22 | This code adds to the input image (`img`) a randomly generated perturbation (`vec1`) which is subjected to a max norm constraint `eps`. This adversarial image lies on a hypercube centerd around the original image. To explore a region (a hypersphere) around the adversarial image (`img + vec1`), we add to it another perturbation (`vec2`) which is constrained by L<sub>2</sub> norm `rad`.  
23 | Pressing keys `e` and `r` generates new `vec1` and `vec2` respectively.  
24 | 
25 | 
26 | 
27 | 
28 | ## Random Perturbations   
29 |  
30 |  The classifier is robust to these random perturbations even though they have severely degraded the image. Perturbations are clearly noticeable and have significantly higher max norm.  
31 |  
32 |  | ![horse_explore](images/horse_explore_single.gif) | ![automobile_explore](images/automobile_explore.gif) | ![truck_explore](images/truck_explore.gif) |  
33 |  |:------------------------------------------:|:-----------------------:|:-----------:|  
34 |  |             **horse**                      |      **automobile**     |: **truck** :|  
35 |  
36 |  In above images, there is no change in class labels and very small drops in probability.
37 | 
38 | 
39 | 
40 | 
41 | ## FGSM Perturbations  
42 | A properly directed perturbation with max norm as low as 3, which is almost imperceptible, can fool the classifier.    
43 | 
44 |  | ![horse_scaled](images/horse_scaled.png) | ![horse_adversarial](images/horse_fgsm.png) | ![perturbation](images/horse_fgsm_pert.png) |
45 |  |:---------:|:--------------------:|:--------------------------:|
46 |  | **horse** |  predicted - **dog** | perturbation **(eps = 6)** |  
47 |  
48 | 
49 | 


--------------------------------------------------------------------------------
/adversarial-examples/random_perturbations/explore_space.py:
--------------------------------------------------------------------------------
  1 | """ Generate random perturbations.
  2 | There are two random vectors here (tensors of shape (32, 32, 3)),
  3 |     vec1 - max norm, eps
  4 |     vec2 - L2 norm, rad
  5 | 
  6 |     vec2 lies on a unit hypersphere and vec1 on hypercube
  7 | 
  8 | Controls:
  9 |     'r' - generate new vec2
 10 |     'e' - generate new vec1
 11 | 
 12 | Basically, perturbation vec1 is added to input image to check how robust the classifier is.
 13 | And to explore a small region around the current adversarial image, (you can imagine them to be
 14 | vectors in 32*32*4 dimensional space) we add vec2, a random vector inside a unit hypersphere. Radius can
 15 | be increased by changing rad. Press 'r' and 'e' for changing vec2 and vec1 respectively.
 16 | 
 17 | From 'Explaining and Harnessing Adversarial Examples' - https://arxiv.org/abs/1412.6572,
 18 |     '''
 19 |     The direction of perturbation, rather than the specific point in space, matters most.
 20 |     Space is not full of pockets of adversarial examples that finely tile the reals like the rational numbers.
 21 |     '''
 22 | This code is to test this.
 23 | 
 24 | """
 25 | import numpy as np
 26 | import cv2
 27 | from torch.autograd import Variable
 28 | import argparse
 29 | import torch
 30 | from models import BasicCNN
 31 | 
 32 | np.random.seed(0)
 33 | 
 34 | cifar10_class_names = {0: 'airplane', 1: 'automobile', 2: 'bird', 3: 'cat', 4: 'deer', 5: 'dog', 6: 'frog', 7: 'horse', 8: 'ship', 9: 'truck'}
 35 | 
 36 | parser = argparse.ArgumentParser()
 37 | parser.add_argument('--img', type=str, default='images/horse.png', help='path to image')
 38 | 
 39 | args = parser.parse_args()
 40 | image_path = args.img
 41 | 
 42 | def random_vector_surface(shape=(32, 32, 3)):
 43 |     # generates a random vector on the surface of hypersphere
 44 |     mat = np.random.normal(size=shape)
 45 |     norm = np.linalg.norm(mat)
 46 |     return mat/norm
 47 | 
 48 | def random_vector_volume(shape=(32, 32, 3)):
 49 |     # generates a random vector in the volume of unit hypersphere
 50 |     d = np.random.rand() ** (1 / np.prod(shape))
 51 | 
 52 |     return random_vector_surface() * d
 53 | 
 54 | 
 55 | window_pert = 'perturbation'
 56 | cv2.namedWindow(window_pert)
 57 | 
 58 | 
 59 | eps, rad = 0, 0
 60 | 
 61 | def get_radius(x):
 62 |     global eps, rad
 63 |     eps = cv2.getTrackbarPos('eps', window_pert)
 64 |     rad = cv2.getTrackbarPos('radius', window_pert)
 65 | 
 66 | 
 67 | cv2.createTrackbar('eps', window_pert, 1, 255, get_radius)
 68 | cv2.createTrackbar('radius', window_pert, 0, 255, get_radius)
 69 | orig = cv2.imread(image_path)[..., ::-1] # BGR -> RGB
 70 | orig = cv2.resize(orig, (32, 32))
 71 | 
 72 | 
 73 | vec1 = random_vector_surface()
 74 | vec2 = random_vector_volume()
 75 | pert = np.zeros((32, 32, 3), dtype=np.float32)
 76 | 
 77 | 
 78 | model = BasicCNN()
 79 | saved = torch.load("cifar10_basiccnn.pth.tar", map_location='cpu')
 80 | model.load_state_dict(saved['state_dict'])
 81 | model.eval()
 82 | 
 83 | img = orig.astype(np.float32)/255.0
 84 | img = img.transpose(2, 0, 1)
 85 | 
 86 | def scale(x, scale=10):
 87 |     return cv2.resize(x, None, fx=scale, fy=scale, interpolation=cv2.INTER_AREA)
 88 | 
 89 | def softmax(x):
 90 |     e_x = np.exp(x - np.max(x))
 91 |     return e_x / e_x.sum()
 92 | 
 93 | 
 94 | while True:
 95 | 
 96 |     key = cv2.waitKey(50) & 0xFF
 97 |     if key == 27:
 98 |         break
 99 | 
100 |     elif key == ord('e'):
101 |         vec1 = random_vector_surface()
102 | 
103 |     elif key == ord('r'):
104 |         vec2 = random_vector_volume()
105 | 
106 |     pert = (eps/255.0) * np.sign(vec1) + (rad/255.0) * vec2
107 | 
108 |     inp = torch.from_numpy(img).float().unsqueeze(0)
109 | 
110 |     prob = softmax(model(inp).data.numpy())[0]
111 |     pred = np.argmax(prob)
112 | 
113 |     # add perturbation to image
114 |     inp = torch.clamp(inp + torch.from_numpy(pert.transpose(2, 0, 1)).float().unsqueeze(0), min=0, max=1)
115 | 
116 |     # predict on the adversarial image
117 |     prob_adv = softmax(model(inp).data.numpy())[0]
118 |     pred_adv = np.argmax(prob_adv)
119 | 
120 |     print("%s [%f] ---> %s [%f]" %(cifar10_class_names[pred], prob[pred], cifar10_class_names[pred_adv], prob_adv[pred_adv]))
121 |     print()
122 | 
123 |     adv = inp.numpy()[0]
124 |     adv = adv.transpose(1, 2, 0)
125 | 
126 |     adv = adv * 255.0
127 |     adv = adv[..., ::-1] # RGB to BGR
128 |     adv = np.clip(adv, 0, 255).astype(np.uint8)
129 | 
130 |     cv2.imshow(window_pert, scale(pert))
131 |     cv2.imshow('vec1', scale((eps/255.0)*np.sign(vec1)))
132 |     cv2.imshow('vec2', (rad/255.0)*scale(vec2))
133 |     cv2.imshow('orig', scale(adv))
134 | cv2.destroyAllWindows()
135 | 


--------------------------------------------------------------------------------
/adversarial-examples/random_perturbations/models.py:
--------------------------------------------------------------------------------
  1 | """
  2 |     Custom models for training on cifar10 and mnist
  3 | 
  4 |     BasicCNN and BasicNN
  5 | """
  6 | 
  7 | import torch.nn.functional as F
  8 | import torch.nn as nn
  9 | 
 10 | class BasicCNN(nn.Module):
 11 |     def __init__(self):
 12 |         super(BasicCNN, self).__init__()
 13 |         """
 14 |             input   - (3, 32, 32)
 15 |             block 1 - (32, 32, 32)
 16 |             maxpool - (32, 16, 16)
 17 |             block 2 - (64, 16, 16)
 18 |             maxpool - (64, 8, 8)
 19 |             block 3 - (128, 8, 8)
 20 |             maxpool - (128, 4, 4)
 21 |             block 4 - (128, 4, 4)
 22 |             avgpool - (128, 1, 1), reshpe to (128,)
 23 |             fc      - (128,) -> (10,)
 24 | 
 25 |         """
 26 |         # block 1
 27 |         self.conv1 = nn.Conv2d(3, 32, kernel_size=3, padding=1)
 28 |         self.conv2 = nn.Conv2d(32, 32, kernel_size=3, padding=1)
 29 |         self.bn1 = nn.BatchNorm2d(32)
 30 | 
 31 |         # block 2
 32 |         self.conv3 = nn.Conv2d(32, 64, kernel_size=3, padding=1)
 33 |         self.conv4 = nn.Conv2d(64, 64, kernel_size=3, padding=1)
 34 |         self.bn2 = nn.BatchNorm2d(64)
 35 | 
 36 |         # block 3
 37 |         self.conv5 = nn.Conv2d(64, 128, kernel_size=3, padding=1)
 38 |         self.conv6 = nn.Conv2d(128, 128, kernel_size=3, padding=1)
 39 |         self.bn3 = nn.BatchNorm2d(128)
 40 | 
 41 |         # block 4
 42 |         self.conv7 = nn.Conv2d(128, 256, kernel_size=3, padding=1)
 43 |         self.conv8 = nn.Conv2d(256, 256, kernel_size=3, padding=1)
 44 |         self.bn4 = nn.BatchNorm2d(256)
 45 | 
 46 |         self.avgpool = nn.AdaptiveAvgPool2d(1)
 47 |         self.fc = nn.Linear(256, 10)
 48 | 
 49 |     def forward(self, x):
 50 | 
 51 |         # block 1
 52 |         x = F.relu(self.conv1(x))
 53 |         x = F.relu(self.bn1(self.conv2(x)))
 54 | 
 55 |         # maxpool
 56 |         x = F.max_pool2d(x, 2)
 57 | 
 58 |         # block 2
 59 |         x = F.relu(self.conv3(x))
 60 |         x = F.relu(self.bn2(self.conv4(x)))
 61 | 
 62 |         # maxpool
 63 |         x = F.max_pool2d(x, 2)
 64 | 
 65 |         # block 3
 66 |         x = F.relu(self.conv5(x))
 67 |         x = F.relu(self.bn3(self.conv6(x)))
 68 | 
 69 |         # maxpool
 70 |         x = F.max_pool2d(x, 2)
 71 | 
 72 |         # block 4
 73 |         x = F.relu(self.conv7(x))
 74 |         x = F.relu(self.bn4(self.conv8(x)))
 75 | 
 76 |         # avgpool and reshape to 1D
 77 |         x = self.avgpool(x)
 78 |         x = x.view(x.size(0), -1)
 79 | 
 80 |         # fc
 81 |         x = self.fc(x)
 82 | 
 83 |         return x
 84 | 
 85 | class BasicNN(nn.Module):
 86 |     def __init__(self):
 87 |         super(BasicNN, self).__init__()
 88 | 
 89 |         self.fc1 = nn.Linear(28*28, 512)
 90 |         self.bn1 = nn.BatchNorm1d(512)
 91 | 
 92 |         self.fc2 = nn.Linear(512, 512)
 93 |         self.bn2 = nn.BatchNorm1d(512)
 94 | 
 95 |         self.fc3 = nn.Linear(512, 256)
 96 |         self.bn3 = nn.BatchNorm1d(256)
 97 | 
 98 | 
 99 |         self.fc4 = nn.Linear(256, 128)
100 |         self.bn4 = nn.BatchNorm1d(128)
101 | 
102 |         self.fc5 = nn.Linear(128, 64)
103 |         self.bn5 = nn.BatchNorm1d(64)
104 | 
105 |         self.fc6 = nn.Linear(64, 10)
106 | 
107 |     def forward(self, x):
108 |         x = F.relu(self.bn1(self.fc1(x)))
109 | 
110 |         x = F.relu(self.bn2(self.fc2(x)))
111 |         x = F.relu(self.bn3(self.fc3(x)))
112 | 
113 |         x = F.relu(self.bn4(self.fc4(x)))
114 |         x = F.relu(self.bn5(self.fc5(x)))
115 | 
116 |         x = self.fc6(x)
117 | 
118 |         return x
119 | 


--------------------------------------------------------------------------------
/adversarial-examples/spatially_transformed/README.md:
--------------------------------------------------------------------------------
 1 | # Spatially Transformed Adversarial Examples
 2 | [Paper](https://arxiv.org/abs/1801.02612) | ICLR 2018  
 3 | For clarity refer [View Synthesis by Appearance Flow](https://people.eecs.berkeley.edu/~tinghuiz/papers/eccv16_appflow.pdf).
 4 | 
 5 | 
 6 | ## Usage
 7 | ```bash
 8 | $ python3 stadv.py --img images/1.jpg --target 7
 9 | ```  
10 | Requires OpenCV for real-time visualization.  
11 | 
12 | 
13 | ## Demo
14 | ![0_1](images/demo/0_1.gif) ![1_2](images/demo/1_2.gif) ![2_3](images/demo/2_3.gif) ![3_4](images/demo/3_4.gif) ![4_5](images/demo/4_5.gif) ![5_6](images/demo/5_6.gif) ![6_7](images/demo/6_7.gif) ![7_8](images/demo/7_8.gif) ![8_9](images/demo/8_9.gif) ![9_0](images/demo/9_0.gif)
15 | 
16 | ## Results  
17 | #### MNIST
18 | Column index is target label and ground truth images are along diagonal. 
19 |   
20 |   
21 | ![tile](images/tile.png?raw=true)
22 | 
23 | 


--------------------------------------------------------------------------------
/adversarial-examples/spatially_transformed/model_mnist.py:
--------------------------------------------------------------------------------
 1 | import torch.nn as nn
 2 | import torch
 3 | import torch.nn.functional as F
 4 | 
 5 | 
 6 | class Basic_CNN(nn.Module):
 7 |     def __init__(self, in_channels, num_classes):
 8 |         super(Basic_CNN, self).__init__()
 9 | 
10 |         self.in_channels = in_channels
11 |         self.num_classes = num_classes
12 | 
13 |         self.conv1_1 = nn.Conv2d(self.in_channels, 32, kernel_size=3, padding=1)
14 |         self.conv1_2 = nn.Conv2d(32, 32, kernel_size=3, padding=1)
15 | 
16 |         self.maxpool1 = nn.MaxPool2d(kernel_size=2)
17 | 
18 |         self.conv2_1 = nn.Conv2d(32, 64, kernel_size=3, padding=1)
19 |         self.conv2_2 = nn.Conv2d(64, 64, kernel_size=3, padding=1)
20 | 
21 |         self.maxpool2 = nn.MaxPool2d(kernel_size=2)
22 | 
23 |         self.fc1 = nn.Linear(7*7*64, 200)
24 |         self.fc2 = nn.Linear(200, self.num_classes)
25 | 
26 |     def forward(self, x):
27 | 
28 |         x = F.relu(self.conv1_1(x))
29 |         x = F.relu(self.conv1_2(x))
30 | 
31 |         x = self.maxpool1(x)
32 | 
33 |         x = F.relu(self.conv2_1(x))
34 |         x = F.relu(self.conv2_2(x))
35 | 
36 |         x = self.maxpool2(x)
37 | 
38 |         x = x.view(x.size(0), -1)
39 |         x = F.relu(self.fc1(x))
40 |         x = self.fc2(x)
41 | 
42 |         return x
43 | 
44 | 
45 | if __name__ == '__main__':
46 |     saved = torch.load('9920.pth.tar', map_location='cpu')
47 |     model = Basic_CNN(1, 10)
48 | 
49 |     model.load_state_dict(saved['state_dict'])
50 | 


--------------------------------------------------------------------------------
/adversarial-examples/spatially_transformed/stadv.py:
--------------------------------------------------------------------------------
  1 | """ Spatially Transformed Adversarial Examples
  2 |     Paper link: https://arxiv.org/abs/1801.02612
  3 | """
  4 | import torch
  5 | from torch.autograd import Variable
  6 | import torch.nn as nn
  7 | import torch.nn.functional as F
  8 | 
  9 | import numpy as np
 10 | import cv2
 11 | import argparse
 12 | from model_mnist import Basic_CNN
 13 | 
 14 | 
 15 | def CWLoss(logits, target, kappa=0):
 16 |     # inputs to the softmax function are called logits.
 17 |     # https://arxiv.org/pdf/1608.04644.pdf
 18 |     target = torch.ones(logits.size(0)).type(logits.type()).fill_(target)
 19 |     target_one_hot = torch.eye(10).type(logits.type())[target.long()]
 20 | 
 21 |     # workaround here.
 22 |     # subtract large value from target class to find other max value
 23 |     # https://github.com/carlini/nn_robust_attacks/blob/master/l2_attack.py
 24 |     real = torch.sum(target_one_hot*logits, 1)
 25 |     other = torch.max((1-target_one_hot)*logits - (target_one_hot*10000), 1)[0]
 26 |     kappa = torch.zeros_like(other).fill_(kappa)
 27 | 
 28 |     return torch.sum(torch.max(other-real, kappa))
 29 | 
 30 | 
 31 | class Loss_flow(nn.Module):
 32 |     def __init__(self, neighbours=np.array([[1, 1, 1], [1, 0, 1], [1, 1, 1]])):
 33 |         super(Loss_flow, self).__init__()
 34 | 
 35 |         filters = []
 36 |         for i in range(neighbours.shape[0]):
 37 |             for j in range(neighbours.shape[1]):
 38 |                 if neighbours[i][j] == 1:
 39 |                     filter = np.zeros((1, neighbours.shape[0], neighbours.shape[1]))
 40 |                     filter[0][i][j] = -1
 41 |                     filter[0][neighbours.shape[0]//2][neighbours.shape[1]//2] = 1
 42 |                     filters.append(filter)
 43 | 
 44 |         filters = np.array(filters)
 45 |         self.filters = torch.from_numpy(filters).float()
 46 | 
 47 |     def forward(self, f):
 48 |         # TODO: padding
 49 |         '''
 50 |         f - f.size() =  [1, h, w, 2]
 51 |             f[0, :, :, 0] - u channel
 52 |             f[0, :, :, 1] - v channel
 53 |         '''
 54 |         f_u = f[:, :, :, 0].unsqueeze(1)
 55 |         f_v = f[:, :, :, 1].unsqueeze(1)
 56 | 
 57 |         diff_u = F.conv2d(f_u, self.filters)[0][0] # don't use squeeze
 58 |         diff_u_sq = torch.mul(diff_u, diff_u)
 59 | 
 60 |         diff_v = F.conv2d(f_v, self.filters)[0][0] # don't use squeeze
 61 |         diff_v_sq = torch.mul(diff_v, diff_v)
 62 | 
 63 |         dist = torch.sqrt(torch.sum(diff_u_sq, dim=0) + torch.sum(diff_v_sq, dim=0))
 64 |         return torch.sum(dist)
 65 | 
 66 | 
 67 | if __name__ == '__main__':
 68 | 
 69 |     parser = argparse.ArgumentParser()
 70 |     parser.add_argument('--img', type=str, default='images/1.jpg', help='path to image')
 71 |     parser.add_argument('--target', type=int, required=True, help='Target label')
 72 |     parser.add_argument('--gpu', action="store_true", default=False)
 73 |     parser.add_argument('--tau', type=float, required=False, default=10, help='balance flow loss')
 74 |     parser.add_argument('--lr', type=float, required=False, default=0.005, help='Learning rate')
 75 | 
 76 |     args = parser.parse_args()
 77 |     img_path = args.img
 78 |     target   = args.target
 79 |     gpu = args.gpu
 80 |     tau = args.tau
 81 |     lr = args.lr
 82 |     IMG_SIZE = 28
 83 |     mean = 0 # for flow initialization
 84 |     std = 0.01
 85 | 
 86 |     print('Spatially Transformed Adversarial Examples')
 87 |     print()
 88 | 
 89 | 
 90 |     orig = cv2.imread(img_path, cv2.IMREAD_GRAYSCALE)
 91 |     img = orig.copy().astype(np.float32)
 92 |     perturbation = np.empty_like(orig)
 93 | 
 94 |     mean = [0.5]
 95 |     std = [0.5]
 96 |     img /= 255.0
 97 |     img = (img - mean)/std
 98 | 
 99 | 
100 |     # load model
101 |     model = Basic_CNN(1, 10)
102 |     saved = torch.load('9920.pth.tar', map_location='cpu')
103 |     model.load_state_dict(saved['state_dict'])
104 |     model.eval()
105 | 
106 | 
107 |     # prediction before attack
108 |     x = Variable(torch.from_numpy(img).float().unsqueeze(0).unsqueeze(0), requires_grad=True)
109 | 
110 |     out = model(x)
111 |     pred = np.argmax(out.data.cpu().numpy())
112 |     print('Prediction before attack: %s' %(pred))
113 | 
114 |     if pred == target:
115 |         print('Prediction is same as target class.')
116 |         exit()
117 | 
118 | 
119 |     # flow, grid, loss_functions
120 |     theta = torch.tensor([[1, 0, 0], [0, 1, 0]]).unsqueeze(0).float() # identity transformation
121 |     grid = F.affine_grid(theta, x.size()) # flow = 0. This is base grid
122 |     # grid.size() = (1, h, w, 2)
123 | 
124 |     f = Variable(torch.zeros_like(grid).float(), requires_grad=True)
125 |     torch.nn.init.normal_(f, mean=0, std=0.01)
126 | 
127 |     grid_new = grid + f
128 |     grid_new = grid_new.clamp(min=-1, max=1)
129 |     x_new = F.grid_sample(x, grid_new, mode='bilinear')
130 | 
131 | 
132 |     optimizer = torch.optim.SGD([f,], lr=lr) # optimizer = torch.optim.LBFGS([f, ], lr=lr)
133 | 
134 |     loss_flow = Loss_flow()
135 |     loss_adv = CWLoss
136 | 
137 |     i=0
138 |     while True:
139 |         optimizer.zero_grad()
140 | 
141 |         logits = model(x_new) # .detach() for LBFGS
142 |         pred = np.argmax(logits.data.numpy())
143 | 
144 |         loss = loss_adv(logits, target) + tau*loss_flow(f)
145 |         loss.backward()
146 |         optimizer.step()
147 | 
148 |         # update variables and predict on adversarial image
149 |         grid_new = grid + f
150 |         grid_new = grid_new.clamp(min=-1, max=1)
151 |         x_new = F.grid_sample(x, grid_new, mode='bilinear')
152 | 
153 |         pred_adv = np.argmax(model(x_new).data.numpy())
154 | 
155 |         i+=1
156 |         print("step %d: [%d] \t" %(i, pred_adv))
157 | 
158 | 
159 |         adv = x_new.data[0][0]
160 |         adv = np.clip(adv.numpy(), -1, 1)
161 |         adv = (adv * 0.5 + 0.5)*255
162 |         adv = adv.astype(np.uint8)
163 | 
164 |         cv2.imshow('adv', adv)
165 |         cv2.imshow('orig', orig)
166 |         key = cv2.waitKey(500) & 0xFF
167 |         key2 = 0
168 |         if key == 32:
169 |             while True:
170 |                 key2 = cv2.waitKey(100) & 0xFF
171 |                 if key2 == 32 or key2 == 27:
172 |                     break
173 |                 if key2 == ord('s'):
174 |                     cv2.imwrite('adv.png', adv)
175 |                     cv2.imwrite('orig.png', orig)
176 |         if pred_adv == target:
177 |             while True:
178 |                 key2 = cv2.waitKey(100) & 0xFF
179 |                 if key2 == 32 or key2 == 27:
180 |                     break
181 |                 if key2 == ord('s'):
182 |                     cv2.imwrite('images/results/%d_%d.png'%(9, target), adv)
183 | 
184 |         if key == 27 or key2 == 27:
185 |             break
186 |     print()
187 |     cv2.destroyAllWindows()
188 | 


--------------------------------------------------------------------------------
/code/active.py:
--------------------------------------------------------------------------------
 1 | import random
 2 | import math
 3 | 
 4 | 
 5 | def random_active(candidate_dataset, aim_dataset, active_ratio, totality):
 6 |     length = len(candidate_dataset.samples)
 7 |     print(length)
 8 |     index = random.sample(range(length), round(totality * active_ratio))
 9 |     print(index)
10 |     aim_dataset.add_item(candidate_dataset.samples[index])
11 |     candidate_dataset.remove_item(index)
12 | 
13 | 
14 | def uncertainty_active(candidate_dataset, aim_dataset, uncertainty_rank, current_acc, active_ratio, totality):
15 |     length = len(uncertainty_rank)
16 |     num_active = math.ceil(totality * active_ratio)
17 | 
18 |     print('current_acc: {}'.format(current_acc))
19 |     start = round(current_acc * length)
20 |     if length - start < num_active:
21 |         start = length - num_active
22 |     index = random.sample(range(start, length), num_active)
23 |     print('range = {},  {}'.format(start, length))
24 |     print(index)
25 | 
26 |     active_samples = uncertainty_rank[index, 0:2, ...]
27 |     candidate_ds_index = uncertainty_rank[index, 2, ...]
28 | 
29 |     aim_dataset.add_item(active_samples)
30 |     candidate_dataset.remove_item(candidate_ds_index)
31 | 
32 |     return active_samples
33 | 


--------------------------------------------------------------------------------
/code/image_list.py:
--------------------------------------------------------------------------------
 1 | from torchvision.datasets import VisionDataset
 2 | import warnings
 3 | import torch
 4 | from PIL import Image
 5 | import os
 6 | import os.path
 7 | import numpy as np
 8 | 
 9 | 
10 | def pil_loader(path):
11 |     # open path as file to avoid ResourceWarning (https://github.com/python-pillow/Pillow/issues/835)
12 |     with open(path, 'rb') as f:
13 |         img = Image.open(f)
14 |         return img.convert('RGB')
15 | 
16 | 
17 | class ImageList(VisionDataset):
18 |     """
19 |     Args:
20 |         root (string): Root directory of dataset
21 |         transform (callable, optional): A function/transform that  takes in an PIL image
22 |             and returns a transformed version. E.g, ``transforms.RandomCrop``
23 |         target_transform (callable, optional): A function/transform that takes in the
24 |             target and transforms it.
25 |     """
26 | 
27 |     def __init__(self, root, transform=None, target_transform=None):
28 |         super(ImageList, self).__init__(root, transform=transform, target_transform=target_transform)
29 | 
30 |         # self.samples = np.loadtxt(root, dtype=np.unicode_, delimiter=' ')
31 |         self.samples = np.loadtxt(root, dtype=np.dtype((np.unicode_, 1000)), delimiter=' ')
32 |         self.loader = pil_loader
33 | 
34 |     def __getitem__(self, index):
35 | 
36 |         path, target = self.samples[index]
37 |         target = int(target)
38 | 
39 |         sample = self.loader(path)
40 | 
41 |         if self.transform is not None:
42 |             sample = self.transform(sample)
43 |         if self.target_transform is not None:
44 |             target = self.target_transform(target)
45 |         return sample, target, path
46 | 
47 |     def __len__(self):
48 |         return len(self.samples)
49 | 
50 |     def add_item(self, addition):
51 |         self.samples = np.concatenate((self.samples, addition), axis=0)
52 |         return self.samples
53 | 
54 |     def remove_item(self, reduced):
55 |         self.samples = np.delete(self.samples, reduced, axis=0)
56 |         return self.samples
57 | 


--------------------------------------------------------------------------------
/code/main.py:
--------------------------------------------------------------------------------
  1 | from __future__ import print_function
  2 | import argparse
  3 | import torch
  4 | import torch.nn as nn
  5 | import torch.nn.functional as F
  6 | import torch.optim as optim
  7 | from torch.utils.data import DataLoader
  8 | import torch.backends.cudnn as cudnn
  9 | from torchvision import datasets, transforms
 10 | from image_list import ImageList
 11 | from prepocess import *
 12 | from network import ResNet50Fc, Discriminator, MultiClassify
 13 | from active import random_active, uncertainty_active
 14 | import numpy as np
 15 | from tensorboardX import SummaryWriter
 16 | from utils import single_entropy, margin, get_consistency
 17 | 
 18 | import random
 19 | 
 20 | 
 21 | def train(args, model, device, train_loader, optimizer, epoch):
 22 |     model.train()
 23 |     for batch_idx, (data, target, path) in enumerate(train_loader):
 24 |         data, target = data.to(device), target.to(device)
 25 |         optimizer.zero_grad()
 26 |         feature, output = model(data)
 27 |         loss = F.cross_entropy(output, target)
 28 |         loss.backward()
 29 |         optimizer.step()
 30 |         if batch_idx % args.log_interval == 0:
 31 |             print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
 32 |                 epoch, batch_idx * len(data), len(train_loader.dataset),
 33 |                        100. * batch_idx / len(train_loader), loss.item()))
 34 | 
 35 | 
 36 | def train_sim(args, model, discrim, device, source_train_loader, target_train_loader, optimizer, epoch):
 37 |     model.eval()
 38 |     discrim.train()
 39 |     for batch_idx, ((source_data, source_label, source_path), (target_data, target_label, target_path)) in enumerate(
 40 |             zip(source_train_loader, target_train_loader)):
 41 |         source_data, source_label = source_data.to(device), source_label.to(device)
 42 |         target_data, target_label = target_data.to(device), target_label.to(device)
 43 | 
 44 |         optimizer.zero_grad()
 45 | 
 46 |         with torch.no_grad():
 47 |             source_feature, source_output = model(source_data)
 48 |             target_feature, target_output = model(target_data)
 49 | 
 50 |         source_sim = discrim(source_feature.detach())
 51 |         target_sim = discrim(target_feature.detach())
 52 | 
 53 |         sim_loss = F.binary_cross_entropy(source_sim, torch.zeros_like(source_sim)) + \
 54 |                    F.binary_cross_entropy(target_sim, torch.ones_like(target_sim))
 55 |         sim_loss.backward()
 56 |         optimizer.step()
 57 | 
 58 |         if batch_idx % args.log_interval == 0:
 59 |             print('Sim Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
 60 |                 epoch, batch_idx * len(source_data),
 61 |                 min(len(source_train_loader.dataset), len(target_train_loader.dataset)),
 62 |                        100. * batch_idx / min(len(source_train_loader), len(target_train_loader)), sim_loss.item()))
 63 | 
 64 | 
 65 | def train_multi(args, model, model1, device, train_loader1, train_loader2, train_loader3, train_loader4, train_loader5,
 66 |                 optimizer1, epoch):
 67 |     model.eval()
 68 |     model1.train()
 69 |     iters = zip(train_loader1, train_loader2, train_loader3, train_loader4, train_loader5)
 70 | 
 71 |     for batch_idx, ((data1, target1, path1), (data2, target2, path2), (data3, target3, path3),
 72 |                     (data4, target4, path4), (data5, target5, path5)) in enumerate(iters):
 73 |         data1 = data1.to(device)
 74 |         data2 = data2.to(device)
 75 |         data3 = data3.to(device)
 76 |         data4 = data4.to(device)
 77 |         data5 = data5.to(device)
 78 | 
 79 |         target1 = target1.to(device)
 80 |         target2 = target2.to(device)
 81 |         target3 = target3.to(device)
 82 |         target4 = target4.to(device)
 83 |         target5 = target5.to(device)
 84 | 
 85 |         with torch.no_grad():
 86 |             feature1, output1 = model(data1)
 87 |             feature2, output2 = model(data2)
 88 |             feature3, output3 = model(data3)
 89 |             feature4, output4 = model(data4)
 90 |             feature5, output5 = model(data5)
 91 | 
 92 |         optimizer1.zero_grad()
 93 | 
 94 |         y1_d1, y2_d1, y3_d1, y4_s1, y5_s1 = model1(feature1.detach())
 95 |         y1_d2, y2_d2, y3_d2, y4_s2, y5_s2 = model1(feature2.detach())
 96 |         y1_d3, y2_d3, y3_d3, y4_s3, y5_s3 = model1(feature3.detach())
 97 |         y1_d4, y2_d4, y3_d4, y4_s4, y5_s4 = model1(feature4.detach())
 98 |         y1_d5, y2_d5, y3_d5, y4_s5, y5_s5 = model1(feature5.detach())
 99 | 
100 |         loss1 = F.cross_entropy(y1_d1, target1)
101 |         loss2 = F.cross_entropy(y2_d2, target2)
102 |         loss3 = F.cross_entropy(y3_d3, target3)
103 |         loss4 = F.cross_entropy(y4_s4, target4)
104 |         loss5 = F.cross_entropy(y5_s5, target5)
105 |         loss = loss1 + loss2 + loss3 + loss4 + loss5
106 | 
107 |         loss.backward()
108 |         optimizer1.step()
109 | 
110 | 
111 | def test(args, model, device, test_loader, multi):
112 |     model.eval()
113 |     multi.eval()
114 |     test_loss = 0
115 |     correct = 0
116 |     with torch.no_grad():
117 |         for data, target, path in test_loader:
118 |             data, target = data.to(device), target.to(device)
119 |             feature, output = model(data)
120 |             test_loss += F.cross_entropy(output, target, reduction='sum').item()  # sum up batch loss
121 |             pred = output.argmax(dim=1, keepdim=True)  # get the index of the max log-probability
122 |             correct += pred.eq(target.view_as(pred)).sum().item()
123 | 
124 |     test_loss /= len(test_loader.dataset)
125 | 
126 |     print('\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.2f}%)\n'.format(
127 |         test_loss, correct, len(test_loader.dataset),
128 |         100. * correct / len(test_loader.dataset)))
129 | 
130 |     return correct / len(test_loader.dataset)
131 | 
132 | 
133 | def find(args, model, model1, device, train_loader):
134 |     model.eval()
135 |     model1.eval()
136 |     stat = list()
137 |     with torch.no_grad():
138 |         for batch_idx, (data, target, path) in enumerate(train_loader):
139 |             data, target = data.to(device), target.to(device)
140 |             feature, output = model(data)
141 |             target_sim = model1(feature.detach())
142 | 
143 |             entropy = single_entropy(output)
144 | 
145 |             pred = output.argmax(dim=1, keepdim=True)  # get the index of the max log-probability
146 |             correct = pred.eq(target.view_as(pred))
147 | 
148 |             for i in range(len(correct)):
149 |                 stat.append([path[i], target[i].item(), batch_idx * args.batch_size + i,
150 |                              pred[i].item(), entropy[i].item(), correct[i].item()])
151 | 
152 |     stat = sorted(stat, key=lambda x: x[0])
153 |     np.savetxt('stat.csv', stat, delimiter=',', fmt='%s')
154 |     return stat
155 | 
156 | 
157 | def uncertainty_evaluate(args, model, multi, discrim, device, train_loader):
158 |     model.eval()
159 |     multi.eval()
160 |     stat = list()
161 |     with torch.no_grad():
162 |         for batch_idx, (data, target, path) in enumerate(train_loader):
163 |             data, target = data.to(device), target.to(device)
164 |             feature, output = model(data)
165 |             y1, y2, y3, y4, y5 = multi(feature)
166 |             target_sim = discrim(feature.detach())
167 | 
168 |             # uncertainty = margin(output) + get_consistency(y1, y2, y3, y4, y5) + target_sim
169 |             uncertainty = margin(output) + get_consistency(y1, y2, y3, y4, y5)
170 |             pred = output.argmax(dim=1, keepdim=True)  # get the index of the max log-probability
171 |             correct = pred.eq(target.view_as(pred))
172 | 
173 |             for i in range(len(correct)):
174 |                 stat.append([path[i], target[i].item(), batch_idx * args.batch_size + i,
175 |                              pred[i].item(), uncertainty[i].item(), correct[i].item()])
176 | 
177 |     stat = sorted(stat, key=lambda x: x[4])
178 |     stat = np.array(stat)
179 |     return stat
180 | 
181 | 
182 | def main():
183 |     # Training settings
184 |     parser = argparse.ArgumentParser(description='PyTorch Example')
185 |     parser.add_argument('--batch-size', type=int, default=32, metavar='N',
186 |                         help='input batch size for training (default: 32)')
187 |     parser.add_argument('--test-batch-size', type=int, default=256, metavar='N',
188 |                         help='input batch size for testing (default: 1000)')
189 |     parser.add_argument('--epochs', type=int, default=20, metavar='N',
190 |                         help='number of epochs to train (default: 20)')
191 |     parser.add_argument('--lr', type=float, default=0.1, metavar='LR',
192 |                         help='learning rate (default: 1.0)')
193 |     parser.add_argument('--gamma', type=float, default=0.7, metavar='M',
194 |                         help='Learning rate step gamma (default: 0.7)')
195 |     parser.add_argument('--gpu', default=None, type=str,
196 |                         help='GPU id to use.')
197 |     parser.add_argument('--seed', type=int, default=1, metavar='S',
198 |                         help='random seed (default: 1)')
199 |     parser.add_argument('--log-interval', type=int, default=50, metavar='N',
200 |                         help='how many batches to wait before logging training status')
201 |     parser.add_argument('--save-model', action='store_true', default=False,
202 |                         help='For Saving the current Model')
203 |     parser.add_argument('--source', type=str, default='', help="The source dataset path list")
204 |     parser.add_argument('--source-val', type=str, default='', help="The source validation dataset path list")
205 |     parser.add_argument('--target', type=str, default='', help="The target dataset path list")
206 |     parser.add_argument('--target-val', type=str, default='', help="The target validation dataset path list")
207 |     parser.add_argument('--class-num', default=31, type=int, help='class num of dataset.')
208 |     args = parser.parse_args()
209 |     use_cuda = args.gpu and torch.cuda.is_available()
210 | 
211 |     torch.manual_seed(args.seed)
212 |     torch.cuda.manual_seed(args.seed)
213 |     random.seed(args.seed)
214 |     # cudnn.benchmark = True
215 |     cudnn.deterministic = True
216 |     device = torch.device("cuda:" + args.gpu if use_cuda else "cpu")
217 | 
218 |     writer = SummaryWriter()
219 |     kwargs = {'num_workers': 2, 'pin_memory': True} if use_cuda else {}
220 | 
221 |     source_train_ds = ImageList(args.source, transform=train_transform)
222 |     source_train_ds1 = ImageList(args.source, transform=train_transform1)
223 |     source_train_ds2 = ImageList(args.source, transform=train_transform2)
224 |     source_train_ds3 = ImageList(args.source, transform=train_transform3)
225 |     source_train_ds4 = ImageList(args.source, transform=train_transform4)
226 |     source_train_ds5 = ImageList(args.source, transform=train_transform5)
227 | 
228 |     target_train_ds = ImageList(args.target, transform=test_transform)
229 |     target_val_ds = ImageList(args.target_val, transform=test_transform)
230 | 
231 |     source_train_loader = DataLoader(source_train_ds, batch_size=args.batch_size, shuffle=True, drop_last=True,
232 |                                      **kwargs)
233 |     train_loader1 = DataLoader(source_train_ds1, batch_size=args.batch_size, shuffle=True, drop_last=True, **kwargs)
234 |     train_loader2 = DataLoader(source_train_ds2, batch_size=args.batch_size, shuffle=True, drop_last=True, **kwargs)
235 |     train_loader3 = DataLoader(source_train_ds3, batch_size=args.batch_size, shuffle=True, drop_last=True, **kwargs)
236 |     train_loader4 = DataLoader(source_train_ds4, batch_size=args.batch_size, shuffle=True, drop_last=True, **kwargs)
237 |     train_loader5 = DataLoader(source_train_ds5, batch_size=args.batch_size, shuffle=True, drop_last=True, **kwargs)
238 | 
239 |     target_train_loader = DataLoader(target_train_ds, batch_size=args.batch_size, **kwargs)
240 |     target_test_loader = DataLoader(target_val_ds, batch_size=args.test_batch_size, **kwargs)
241 | 
242 |     model = ResNet50Fc(bottleneck_dim=256, class_num=args.class_num).to(device)
243 |     multi = MultiClassify(bottleneck_dim=256, class_num=args.class_num).to(device)
244 |     discrim = Discriminator(bottleneck_dim=256).to(device)
245 | 
246 |     optimizer = optim.Adadelta(model.parameters_list(args.lr), lr=args.lr)
247 |     optimizer1 = optim.Adadelta(multi.parameters(), lr=args.lr)
248 | 
249 |     totality = len(target_train_ds)
250 |     for epoch in range(1, args.epochs + 1):
251 | 
252 |         train(args, model, device, source_train_loader, optimizer, epoch)
253 |         train_multi(args, model, multi, device, train_loader1, train_loader2, train_loader3, train_loader4,
254 |                     train_loader5, optimizer1, epoch)
255 |         train_sim(args, model, discrim, device, source_train_loader, target_train_loader, optimizer, epoch)
256 |         test_acc = test(args, model, device, target_test_loader, multi)
257 | 
258 |         # print(test_acc)
259 |         writer.add_scalar('testacc', test_acc, epoch)
260 | 
261 |         if epoch in [10, 12, 14, 16, 18]:
262 |             # if epoch in [14, 16, 18, 20, 22, 24, 28, 30, 32, 34]:
263 |             # if epoch in [8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 28, 30, 32, 34, 36, 38, 40, 42, 44, 46]:
264 |             # if epoch in [2, 16]:
265 |             # random_active(candidate_dataset=target_train_ds, aim_dataset=source_train_ds, active_ratio=0.01,
266 |             #               totality=totality)
267 |             uncertainty_rank = uncertainty_evaluate(args, model, multi, discrim, device, target_train_loader)
268 |             active_samples = uncertainty_active(candidate_dataset=target_train_ds, aim_dataset=source_train_ds,
269 |                                                 uncertainty_rank=uncertainty_rank, current_acc=test_acc,
270 |                                                 active_ratio=0.01, totality=totality)
271 |             source_train_ds1.add_item(active_samples)
272 |             source_train_ds2.add_item(active_samples)
273 |             source_train_ds3.add_item(active_samples)
274 |             source_train_ds4.add_item(active_samples)
275 |             source_train_ds5.add_item(active_samples)
276 | 
277 |     # np.savetxt(args.source[17] + '-' + args.target[17] + '.txt', source_train_ds.samples, delimiter=' ', fmt='%s')
278 | 
279 |     # wrong_list = find(args, model, model1, device, target_test_loader)
280 |     # np.savetxt('sim.txt', wrong_list, fmt='%.5f')
281 |     # np.savetxt('wrong_list.txt', wrong_list, fmt='%s')
282 | 
283 |     if args.save_model:
284 |         torch.save(model.state_dict(), "resnet.pt")
285 | 
286 |     # writer.export_scalars_to_json("./all_scalars.json")
287 |     writer.close()
288 | 
289 | 
290 | if __name__ == '__main__':
291 |     main()
292 | 


--------------------------------------------------------------------------------
/code/network.py:
--------------------------------------------------------------------------------
  1 | import torch
  2 | import torch.nn as nn
  3 | import torchvision
  4 | from torchvision import models
  5 | import math
  6 | 
  7 | 
  8 | class ResNet50Fc(nn.Module):
  9 | 
 10 |     def __init__(self, bottleneck_dim=256, class_num=1000):
 11 |         super(ResNet50Fc, self).__init__()
 12 |         self.model_resnet = models.resnet50(pretrained=True)
 13 | 
 14 |         model_resnet = self.model_resnet
 15 |         self.conv1 = model_resnet.conv1
 16 |         self.bn1 = model_resnet.bn1
 17 |         self.relu = model_resnet.relu
 18 |         self.maxpool = model_resnet.maxpool
 19 |         self.layer1 = model_resnet.layer1
 20 |         self.layer2 = model_resnet.layer2
 21 |         self.layer3 = model_resnet.layer3
 22 |         self.layer4 = model_resnet.layer4
 23 |         self.avgpool = model_resnet.avgpool
 24 |         self.bottleneck = nn.Linear(model_resnet.fc.in_features, bottleneck_dim)
 25 |         self.bn2 = nn.BatchNorm1d(bottleneck_dim)
 26 |         self.fc = nn.Linear(bottleneck_dim, class_num)
 27 |         # self.fc = nn.Linear(model_resnet.fc.in_features, class_num)
 28 | 
 29 |     def forward(self, x):
 30 |         x = self.conv1(x)
 31 |         x = self.bn1(x)
 32 |         x = self.relu(x)
 33 |         x = self.maxpool(x)
 34 |         x = self.layer1(x)
 35 |         x = self.layer2(x)
 36 |         x = self.layer3(x)
 37 |         x = self.layer4(x)
 38 |         x = self.avgpool(x)
 39 |         x = torch.flatten(x, 1)
 40 |         x = self.bottleneck(x)
 41 |         x = self.bn2(x)
 42 |         y = self.fc(x)
 43 |         return x, y
 44 | 
 45 |     def output_num(self):
 46 |         return self.__in_features
 47 | 
 48 |     def parameters_list(self, lr):
 49 |         parameter_list = [
 50 |             {'params': self.conv1.parameters(), 'lr': lr / 10},
 51 |             {'params': self.bn1.parameters(), 'lr': lr / 10},
 52 |             {'params': self.maxpool.parameters(), 'lr': lr / 10},
 53 |             {'params': self.layer1.parameters(), 'lr': lr / 10},
 54 |             {'params': self.layer2.parameters(), 'lr': lr / 10},
 55 |             {'params': self.layer3.parameters(), 'lr': lr / 10},
 56 |             {'params': self.layer4.parameters(), 'lr': lr / 10},
 57 |             {'params': self.avgpool.parameters(), 'lr': lr / 10},
 58 |             {'params': self.bottleneck.parameters()},
 59 |             # {'params': self.bn2.parameters()},
 60 |             {'params': self.fc.parameters()},
 61 |         ]
 62 | 
 63 |         return parameter_list
 64 | 
 65 | 
 66 | class Discriminator(nn.Module):
 67 |     def __init__(self, bottleneck_dim=256):
 68 |         super(Discriminator, self).__init__()
 69 |         self.fc = nn.Linear(bottleneck_dim, 1)
 70 |         self.sigmoid = nn.Sigmoid()
 71 | 
 72 |         nn.init.kaiming_uniform_(self.fc.weight)
 73 | 
 74 |     def forward(self, x):
 75 |         x = self.fc(x)
 76 |         x = self.sigmoid(x)
 77 |         # x = torch.flatten(x)
 78 |         return x
 79 | 
 80 |     def parameters_list(self, lr):
 81 |         parameter_list = [
 82 |             {'params': self.fc.parameters()},
 83 |             {'params': self.sigmoid.parameters()}
 84 |         ]
 85 |         return parameter_list
 86 | 
 87 | 
 88 | class MultiClassify(nn.Module):
 89 | 
 90 |     def __init__(self, bottleneck_dim=256, class_num=1000):
 91 |         super(MultiClassify, self).__init__()
 92 | 
 93 |         self.fc1 = nn.Linear(bottleneck_dim, class_num)
 94 |         self.fc2 = nn.Linear(bottleneck_dim, class_num)
 95 |         self.fc3 = nn.Linear(bottleneck_dim, class_num)
 96 |         self.fc4 = nn.Linear(bottleneck_dim, class_num)
 97 |         self.fc5 = nn.Linear(bottleneck_dim, class_num)
 98 | 
 99 |         nn.init.xavier_uniform_(self.fc1.weight, gain=nn.init.calculate_gain('relu'))
100 |         nn.init.xavier_normal_(self.fc2.weight)
101 |         nn.init.kaiming_uniform_(self.fc3.weight)
102 |         nn.init.kaiming_uniform_(self.fc4.weight, nonlinearity='relu')
103 |         nn.init.kaiming_normal_(self.fc5.weight, a=math.sqrt(5))
104 | 
105 |     def forward(self, x):
106 |         y1 = self.fc1(x)
107 |         y2 = self.fc2(x)
108 |         y3 = self.fc3(x)
109 |         y4 = self.fc4(x)
110 |         y5 = self.fc5(x)
111 | 
112 |         return y1, y2, y3, y4, y5
113 | 


--------------------------------------------------------------------------------
/code/prepocess.py:
--------------------------------------------------------------------------------
 1 | from torchvision import datasets, transforms
 2 | from PIL import Image
 3 | 
 4 | train_transform = transforms.Compose([
 5 |     transforms.Resize(256),
 6 |     # transforms.RandomHorizontalFlip(0.5),
 7 |     transforms.RandomCrop(224),
 8 |     transforms.ToTensor(),
 9 |     transforms.Normalize(mean=[0.485, 0.456, 0.406],
10 |                          std=[0.229, 0.224, 0.225]),
11 | ])
12 | 
13 | test_transform = transforms.Compose([
14 |     transforms.Resize(256),
15 |     transforms.CenterCrop(224),
16 |     transforms.ToTensor(),
17 |     transforms.Normalize(mean=[0.485, 0.456, 0.406],
18 |                          std=[0.229, 0.224, 0.225]),
19 | ])
20 | 
21 | train_transform1 = transforms.Compose([
22 |     transforms.Resize(256),
23 |     transforms.RandomHorizontalFlip(),
24 |     transforms.RandomAffine(degrees=30, translate=(0.1, 0.1), scale=(0.9, 1.1), shear=0.2, resample=Image.BICUBIC,
25 |                             fillcolor=(255, 255, 255)),
26 |     transforms.CenterCrop(224),
27 |     transforms.RandomGrayscale(p=0.1),
28 |     transforms.ToTensor(),
29 |     transforms.Normalize(mean=[0.485, 0.456, 0.406],
30 |                          std=[0.229, 0.224, 0.225]),
31 | ])
32 | 
33 | train_transform2 = transforms.Compose([
34 |     transforms.Resize(256),
35 |     transforms.RandomHorizontalFlip(),
36 |     transforms.RandomPerspective(),
37 |     transforms.FiveCrop(224),
38 |     transforms.Lambda(lambda crops: crops[0]),
39 |     transforms.ColorJitter(brightness=0.2, contrast=0.2, saturation=0.2, hue=0.2),
40 |     transforms.ToTensor(),
41 |     transforms.Normalize(mean=[0.485, 0.456, 0.406],
42 |                          std=[0.229, 0.224, 0.225]),
43 | ])
44 | 
45 | train_transform3 = transforms.Compose([
46 |     transforms.Resize(256),
47 |     transforms.RandomHorizontalFlip(),
48 |     transforms.RandomAffine(degrees=30, translate=(0.1, 0.1), scale=(0.9, 1.1), shear=0.2, resample=Image.BICUBIC,
49 |                             fillcolor=(255, 255, 255)),
50 |     transforms.FiveCrop(224),
51 |     transforms.Lambda(lambda crops: crops[1]),
52 |     transforms.ColorJitter(brightness=0.2, contrast=0.2, saturation=0.2, hue=0.2),
53 |     transforms.ToTensor(),
54 |     transforms.Normalize(mean=[0.485, 0.456, 0.406],
55 |                          std=[0.229, 0.224, 0.225]),
56 | ])
57 | 
58 | train_transform4 = transforms.Compose([
59 |     transforms.Resize(256),
60 |     transforms.RandomHorizontalFlip(),
61 |     transforms.RandomAffine(degrees=10, translate=(0.1, 0.1), scale=(0.9, 1.1), shear=0.1, resample=Image.BICUBIC,
62 |                             fillcolor=(255, 255, 255)),
63 |     transforms.RandomPerspective(),
64 |     transforms.FiveCrop(224),
65 |     transforms.Lambda(lambda crops: crops[2]),
66 |     transforms.ToTensor(),
67 |     transforms.Normalize(mean=[0.485, 0.456, 0.406],
68 |                          std=[0.229, 0.224, 0.225]),
69 | ])
70 | 
71 | train_transform5 = transforms.Compose([
72 |     transforms.Resize(256),
73 |     transforms.RandomHorizontalFlip(),
74 |     transforms.RandomPerspective(),
75 |     transforms.FiveCrop(224),
76 |     transforms.Lambda(lambda crops: crops[3]),
77 |     transforms.RandomGrayscale(p=0.1),
78 |     transforms.ToTensor(),
79 |     transforms.Normalize(mean=[0.485, 0.456, 0.406],
80 |                          std=[0.229, 0.224, 0.225]),
81 | ])
82 | 


--------------------------------------------------------------------------------
/code/utils.py:
--------------------------------------------------------------------------------
 1 | import torch
 2 | import torch.nn as nn
 3 | import numpy as np
 4 | 
 5 | 
 6 | def single_entropy(fc2_s):
 7 |     fc2_s = nn.Softmax(-1)(fc2_s)
 8 |     entropy = torch.sum(- fc2_s * torch.log(fc2_s + 1e-10), dim=1)
 9 |     entropy_norm = np.log(fc2_s.size(1))
10 |     entropy = entropy / entropy_norm
11 |     return entropy
12 | 
13 | 
14 | def margin(out):
15 |     out = nn.Softmax(-1)(out)
16 |     top2 = torch.topk(out, 2).values
17 |     # print(top2)
18 |     return 1 - (top2[:, 0] - top2[:, 1])
19 | 
20 | 
21 | def get_entropy(fc2_s, fc2_s2, fc2_s3, fc2_s4, fc2_s5, domain_temperature=1.0, class_temperature=10.0):
22 |     fc2_s = nn.Softmax(-1)(fc2_s)
23 |     fc2_s2 = nn.Softmax(-1)(fc2_s2)
24 |     fc2_s3 = nn.Softmax(-1)(fc2_s3)
25 |     fc2_s4 = nn.Softmax(-1)(fc2_s4)
26 |     fc2_s5 = nn.Softmax(-1)(fc2_s5)
27 | 
28 |     entropy = torch.sum(- fc2_s * torch.log(fc2_s + 1e-10), dim=1)
29 |     entropy2 = torch.sum(- fc2_s2 * torch.log(fc2_s2 + 1e-10), dim=1)
30 |     entropy3 = torch.sum(- fc2_s3 * torch.log(fc2_s3 + 1e-10), dim=1)
31 |     entropy4 = torch.sum(- fc2_s4 * torch.log(fc2_s4 + 1e-10), dim=1)
32 |     entropy5 = torch.sum(- fc2_s5 * torch.log(fc2_s5 + 1e-10), dim=1)
33 |     entropy_norm = np.log(fc2_s.size(1))
34 | 
35 |     weight = (entropy + entropy2 + entropy3 + entropy4 + entropy5) / (5 * entropy_norm)
36 |     return weight
37 | 
38 | 
39 | def get_consistency(fc2_s, fc2_s2, fc2_s3, fc2_s4, fc2_s5):
40 |     fc2_s = nn.Softmax(-1)(fc2_s)
41 |     fc2_s2 = nn.Softmax(-1)(fc2_s2)
42 |     fc2_s3 = nn.Softmax(-1)(fc2_s3)
43 |     fc2_s4 = nn.Softmax(-1)(fc2_s4)
44 |     fc2_s5 = nn.Softmax(-1)(fc2_s5)
45 | 
46 |     fc2_s = torch.unsqueeze(fc2_s, 1)
47 |     fc2_s2 = torch.unsqueeze(fc2_s2, 1)
48 |     fc2_s3 = torch.unsqueeze(fc2_s3, 1)
49 |     fc2_s4 = torch.unsqueeze(fc2_s4, 1)
50 |     fc2_s5 = torch.unsqueeze(fc2_s5, 1)
51 |     c = torch.cat((fc2_s, fc2_s2, fc2_s3, fc2_s4, fc2_s5), dim=1)
52 |     d = torch.std(c, 1)
53 |     consistency = torch.mean(d, 1)
54 |     return consistency
55 | 
56 | 
57 | def get_predict_prob(fc2_s, fc2_s2, fc2_s3, fc2_s4, fc2_s5):
58 |     fc2_s = nn.Softmax(-1)(fc2_s)
59 |     fc2_s2 = nn.Softmax(-1)(fc2_s2)
60 |     fc2_s3 = nn.Softmax(-1)(fc2_s3)
61 |     fc2_s4 = nn.Softmax(-1)(fc2_s4)
62 |     fc2_s5 = nn.Softmax(-1)(fc2_s5)
63 | 
64 |     fc2_s = torch.unsqueeze(fc2_s, 1)
65 |     fc2_s2 = torch.unsqueeze(fc2_s2, 1)
66 |     fc2_s3 = torch.unsqueeze(fc2_s3, 1)
67 |     fc2_s4 = torch.unsqueeze(fc2_s4, 1)
68 |     fc2_s5 = torch.unsqueeze(fc2_s5, 1)
69 |     c = torch.cat((fc2_s, fc2_s2, fc2_s3, fc2_s4, fc2_s5), dim=1)
70 |     predict_prob = torch.mean(c, 1)
71 |     predict_prob = nn.Softmax(-1)(predict_prob)
72 |     return predict_prob
73 | 
74 | 
75 | def get_target_weight(entropy, consistency, threshold):
76 |     sorce = (entropy + consistency) / 2
77 |     weight = [0.0 for i in range(len(sorce))]
78 |     for i in range(len(sorce)):
79 |         if sorce[i] < (threshold / 2):
80 |             weight[i] = 1.0
81 |     return torch.tensor(weight)
82 | 
83 | 
84 | def normalize_weight(x):
85 |     min_val = x.min()
86 |     max_val = x.max()
87 |     x = (x - min_val) / (max_val - min_val)
88 |     return x.detach()
89 | 
90 | 
91 | def nega_normalize_weight(x):
92 |     x = 1 - x
93 |     return x.detach()
94 | 


--------------------------------------------------------------------------------
/data/README.md:
--------------------------------------------------------------------------------
1 | # Active-Domain-Adaptation
2 | 
3 | Dataset download：<br />
4 | Office-31: http://people.eecs.berkeley.edu/~jhoffman/domainadapt/ <br />
5 | Office-Home: http://hemanthdv.org/OfficeHome-Dataset/ <br />
6 | VisDA: https://github.com/VisionLearningGroup/taskcv-2017-public/tree/master/classification <br />
7 | Download the dataset according to the instructions on the above, and update the path in each file in the 'data' folder
8 | 


--------------------------------------------------------------------------------
/data/office/dslr.txt:
--------------------------------------------------------------------------------
  1 | domain_adaptation_images/dslr/images/back_pack/frame_0001.jpg 0
  2 | domain_adaptation_images/dslr/images/back_pack/frame_0002.jpg 0
  3 | domain_adaptation_images/dslr/images/back_pack/frame_0003.jpg 0
  4 | domain_adaptation_images/dslr/images/back_pack/frame_0004.jpg 0
  5 | domain_adaptation_images/dslr/images/back_pack/frame_0005.jpg 0
  6 | domain_adaptation_images/dslr/images/back_pack/frame_0006.jpg 0
  7 | domain_adaptation_images/dslr/images/back_pack/frame_0007.jpg 0
  8 | domain_adaptation_images/dslr/images/back_pack/frame_0008.jpg 0
  9 | domain_adaptation_images/dslr/images/back_pack/frame_0009.jpg 0
 10 | domain_adaptation_images/dslr/images/back_pack/frame_0010.jpg 0
 11 | domain_adaptation_images/dslr/images/back_pack/frame_0011.jpg 0
 12 | domain_adaptation_images/dslr/images/back_pack/frame_0012.jpg 0
 13 | domain_adaptation_images/dslr/images/bike/frame_0001.jpg 1
 14 | domain_adaptation_images/dslr/images/bike/frame_0002.jpg 1
 15 | domain_adaptation_images/dslr/images/bike/frame_0003.jpg 1
 16 | domain_adaptation_images/dslr/images/bike/frame_0004.jpg 1
 17 | domain_adaptation_images/dslr/images/bike/frame_0005.jpg 1
 18 | domain_adaptation_images/dslr/images/bike/frame_0006.jpg 1
 19 | domain_adaptation_images/dslr/images/bike/frame_0007.jpg 1
 20 | domain_adaptation_images/dslr/images/bike/frame_0008.jpg 1
 21 | domain_adaptation_images/dslr/images/bike/frame_0009.jpg 1
 22 | domain_adaptation_images/dslr/images/bike/frame_0010.jpg 1
 23 | domain_adaptation_images/dslr/images/bike/frame_0011.jpg 1
 24 | domain_adaptation_images/dslr/images/bike/frame_0012.jpg 1
 25 | domain_adaptation_images/dslr/images/bike/frame_0013.jpg 1
 26 | domain_adaptation_images/dslr/images/bike/frame_0014.jpg 1
 27 | domain_adaptation_images/dslr/images/bike/frame_0015.jpg 1
 28 | domain_adaptation_images/dslr/images/bike/frame_0016.jpg 1
 29 | domain_adaptation_images/dslr/images/bike/frame_0017.jpg 1
 30 | domain_adaptation_images/dslr/images/bike/frame_0018.jpg 1
 31 | domain_adaptation_images/dslr/images/bike/frame_0019.jpg 1
 32 | domain_adaptation_images/dslr/images/bike/frame_0020.jpg 1
 33 | domain_adaptation_images/dslr/images/bike/frame_0021.jpg 1
 34 | domain_adaptation_images/dslr/images/bike_helmet/frame_0001.jpg 2
 35 | domain_adaptation_images/dslr/images/bike_helmet/frame_0002.jpg 2
 36 | domain_adaptation_images/dslr/images/bike_helmet/frame_0003.jpg 2
 37 | domain_adaptation_images/dslr/images/bike_helmet/frame_0004.jpg 2
 38 | domain_adaptation_images/dslr/images/bike_helmet/frame_0005.jpg 2
 39 | domain_adaptation_images/dslr/images/bike_helmet/frame_0006.jpg 2
 40 | domain_adaptation_images/dslr/images/bike_helmet/frame_0007.jpg 2
 41 | domain_adaptation_images/dslr/images/bike_helmet/frame_0008.jpg 2
 42 | domain_adaptation_images/dslr/images/bike_helmet/frame_0009.jpg 2
 43 | domain_adaptation_images/dslr/images/bike_helmet/frame_0010.jpg 2
 44 | domain_adaptation_images/dslr/images/bike_helmet/frame_0011.jpg 2
 45 | domain_adaptation_images/dslr/images/bike_helmet/frame_0012.jpg 2
 46 | domain_adaptation_images/dslr/images/bike_helmet/frame_0013.jpg 2
 47 | domain_adaptation_images/dslr/images/bike_helmet/frame_0014.jpg 2
 48 | domain_adaptation_images/dslr/images/bike_helmet/frame_0015.jpg 2
 49 | domain_adaptation_images/dslr/images/bike_helmet/frame_0016.jpg 2
 50 | domain_adaptation_images/dslr/images/bike_helmet/frame_0017.jpg 2
 51 | domain_adaptation_images/dslr/images/bike_helmet/frame_0018.jpg 2
 52 | domain_adaptation_images/dslr/images/bike_helmet/frame_0019.jpg 2
 53 | domain_adaptation_images/dslr/images/bike_helmet/frame_0020.jpg 2
 54 | domain_adaptation_images/dslr/images/bike_helmet/frame_0021.jpg 2
 55 | domain_adaptation_images/dslr/images/bike_helmet/frame_0022.jpg 2
 56 | domain_adaptation_images/dslr/images/bike_helmet/frame_0023.jpg 2
 57 | domain_adaptation_images/dslr/images/bike_helmet/frame_0024.jpg 2
 58 | domain_adaptation_images/dslr/images/bookcase/frame_0001.jpg 3
 59 | domain_adaptation_images/dslr/images/bookcase/frame_0002.jpg 3
 60 | domain_adaptation_images/dslr/images/bookcase/frame_0003.jpg 3
 61 | domain_adaptation_images/dslr/images/bookcase/frame_0004.jpg 3
 62 | domain_adaptation_images/dslr/images/bookcase/frame_0005.jpg 3
 63 | domain_adaptation_images/dslr/images/bookcase/frame_0006.jpg 3
 64 | domain_adaptation_images/dslr/images/bookcase/frame_0007.jpg 3
 65 | domain_adaptation_images/dslr/images/bookcase/frame_0008.jpg 3
 66 | domain_adaptation_images/dslr/images/bookcase/frame_0009.jpg 3
 67 | domain_adaptation_images/dslr/images/bookcase/frame_0010.jpg 3
 68 | domain_adaptation_images/dslr/images/bookcase/frame_0011.jpg 3
 69 | domain_adaptation_images/dslr/images/bookcase/frame_0012.jpg 3
 70 | domain_adaptation_images/dslr/images/bottle/frame_0001.jpg 4
 71 | domain_adaptation_images/dslr/images/bottle/frame_0002.jpg 4
 72 | domain_adaptation_images/dslr/images/bottle/frame_0003.jpg 4
 73 | domain_adaptation_images/dslr/images/bottle/frame_0004.jpg 4
 74 | domain_adaptation_images/dslr/images/bottle/frame_0005.jpg 4
 75 | domain_adaptation_images/dslr/images/bottle/frame_0006.jpg 4
 76 | domain_adaptation_images/dslr/images/bottle/frame_0007.jpg 4
 77 | domain_adaptation_images/dslr/images/bottle/frame_0008.jpg 4
 78 | domain_adaptation_images/dslr/images/bottle/frame_0009.jpg 4
 79 | domain_adaptation_images/dslr/images/bottle/frame_0010.jpg 4
 80 | domain_adaptation_images/dslr/images/bottle/frame_0011.jpg 4
 81 | domain_adaptation_images/dslr/images/bottle/frame_0012.jpg 4
 82 | domain_adaptation_images/dslr/images/bottle/frame_0013.jpg 4
 83 | domain_adaptation_images/dslr/images/bottle/frame_0014.jpg 4
 84 | domain_adaptation_images/dslr/images/bottle/frame_0015.jpg 4
 85 | domain_adaptation_images/dslr/images/bottle/frame_0016.jpg 4
 86 | domain_adaptation_images/dslr/images/calculator/frame_0001.jpg 5
 87 | domain_adaptation_images/dslr/images/calculator/frame_0002.jpg 5
 88 | domain_adaptation_images/dslr/images/calculator/frame_0003.jpg 5
 89 | domain_adaptation_images/dslr/images/calculator/frame_0004.jpg 5
 90 | domain_adaptation_images/dslr/images/calculator/frame_0005.jpg 5
 91 | domain_adaptation_images/dslr/images/calculator/frame_0006.jpg 5
 92 | domain_adaptation_images/dslr/images/calculator/frame_0007.jpg 5
 93 | domain_adaptation_images/dslr/images/calculator/frame_0008.jpg 5
 94 | domain_adaptation_images/dslr/images/calculator/frame_0009.jpg 5
 95 | domain_adaptation_images/dslr/images/calculator/frame_0010.jpg 5
 96 | domain_adaptation_images/dslr/images/calculator/frame_0011.jpg 5
 97 | domain_adaptation_images/dslr/images/calculator/frame_0012.jpg 5
 98 | domain_adaptation_images/dslr/images/desk_chair/frame_0001.jpg 6
 99 | domain_adaptation_images/dslr/images/desk_chair/frame_0002.jpg 6
100 | domain_adaptation_images/dslr/images/desk_chair/frame_0003.jpg 6
101 | domain_adaptation_images/dslr/images/desk_chair/frame_0004.jpg 6
102 | domain_adaptation_images/dslr/images/desk_chair/frame_0005.jpg 6
103 | domain_adaptation_images/dslr/images/desk_chair/frame_0006.jpg 6
104 | domain_adaptation_images/dslr/images/desk_chair/frame_0007.jpg 6
105 | domain_adaptation_images/dslr/images/desk_chair/frame_0008.jpg 6
106 | domain_adaptation_images/dslr/images/desk_chair/frame_0009.jpg 6
107 | domain_adaptation_images/dslr/images/desk_chair/frame_0010.jpg 6
108 | domain_adaptation_images/dslr/images/desk_chair/frame_0011.jpg 6
109 | domain_adaptation_images/dslr/images/desk_chair/frame_0012.jpg 6
110 | domain_adaptation_images/dslr/images/desk_chair/frame_0013.jpg 6
111 | domain_adaptation_images/dslr/images/desk_lamp/frame_0001.jpg 7
112 | domain_adaptation_images/dslr/images/desk_lamp/frame_0002.jpg 7
113 | domain_adaptation_images/dslr/images/desk_lamp/frame_0003.jpg 7
114 | domain_adaptation_images/dslr/images/desk_lamp/frame_0004.jpg 7
115 | domain_adaptation_images/dslr/images/desk_lamp/frame_0005.jpg 7
116 | domain_adaptation_images/dslr/images/desk_lamp/frame_0006.jpg 7
117 | domain_adaptation_images/dslr/images/desk_lamp/frame_0007.jpg 7
118 | domain_adaptation_images/dslr/images/desk_lamp/frame_0008.jpg 7
119 | domain_adaptation_images/dslr/images/desk_lamp/frame_0009.jpg 7
120 | domain_adaptation_images/dslr/images/desk_lamp/frame_0010.jpg 7
121 | domain_adaptation_images/dslr/images/desk_lamp/frame_0011.jpg 7
122 | domain_adaptation_images/dslr/images/desk_lamp/frame_0012.jpg 7
123 | domain_adaptation_images/dslr/images/desk_lamp/frame_0013.jpg 7
124 | domain_adaptation_images/dslr/images/desk_lamp/frame_0014.jpg 7
125 | domain_adaptation_images/dslr/images/desktop_computer/frame_0001.jpg 8
126 | domain_adaptation_images/dslr/images/desktop_computer/frame_0002.jpg 8
127 | domain_adaptation_images/dslr/images/desktop_computer/frame_0003.jpg 8
128 | domain_adaptation_images/dslr/images/desktop_computer/frame_0004.jpg 8
129 | domain_adaptation_images/dslr/images/desktop_computer/frame_0005.jpg 8
130 | domain_adaptation_images/dslr/images/desktop_computer/frame_0006.jpg 8
131 | domain_adaptation_images/dslr/images/desktop_computer/frame_0007.jpg 8
132 | domain_adaptation_images/dslr/images/desktop_computer/frame_0008.jpg 8
133 | domain_adaptation_images/dslr/images/desktop_computer/frame_0009.jpg 8
134 | domain_adaptation_images/dslr/images/desktop_computer/frame_0010.jpg 8
135 | domain_adaptation_images/dslr/images/desktop_computer/frame_0011.jpg 8
136 | domain_adaptation_images/dslr/images/desktop_computer/frame_0012.jpg 8
137 | domain_adaptation_images/dslr/images/desktop_computer/frame_0013.jpg 8
138 | domain_adaptation_images/dslr/images/desktop_computer/frame_0014.jpg 8
139 | domain_adaptation_images/dslr/images/desktop_computer/frame_0015.jpg 8
140 | domain_adaptation_images/dslr/images/file_cabinet/frame_0001.jpg 9
141 | domain_adaptation_images/dslr/images/file_cabinet/frame_0002.jpg 9
142 | domain_adaptation_images/dslr/images/file_cabinet/frame_0003.jpg 9
143 | domain_adaptation_images/dslr/images/file_cabinet/frame_0004.jpg 9
144 | domain_adaptation_images/dslr/images/file_cabinet/frame_0005.jpg 9
145 | domain_adaptation_images/dslr/images/file_cabinet/frame_0006.jpg 9
146 | domain_adaptation_images/dslr/images/file_cabinet/frame_0007.jpg 9
147 | domain_adaptation_images/dslr/images/file_cabinet/frame_0008.jpg 9
148 | domain_adaptation_images/dslr/images/file_cabinet/frame_0009.jpg 9
149 | domain_adaptation_images/dslr/images/file_cabinet/frame_0010.jpg 9
150 | domain_adaptation_images/dslr/images/file_cabinet/frame_0011.jpg 9
151 | domain_adaptation_images/dslr/images/file_cabinet/frame_0012.jpg 9
152 | domain_adaptation_images/dslr/images/file_cabinet/frame_0013.jpg 9
153 | domain_adaptation_images/dslr/images/file_cabinet/frame_0014.jpg 9
154 | domain_adaptation_images/dslr/images/file_cabinet/frame_0015.jpg 9
155 | domain_adaptation_images/dslr/images/headphones/frame_0001.jpg 10
156 | domain_adaptation_images/dslr/images/headphones/frame_0002.jpg 10
157 | domain_adaptation_images/dslr/images/headphones/frame_0003.jpg 10
158 | domain_adaptation_images/dslr/images/headphones/frame_0004.jpg 10
159 | domain_adaptation_images/dslr/images/headphones/frame_0005.jpg 10
160 | domain_adaptation_images/dslr/images/headphones/frame_0006.jpg 10
161 | domain_adaptation_images/dslr/images/headphones/frame_0007.jpg 10
162 | domain_adaptation_images/dslr/images/headphones/frame_0008.jpg 10
163 | domain_adaptation_images/dslr/images/headphones/frame_0009.jpg 10
164 | domain_adaptation_images/dslr/images/headphones/frame_0010.jpg 10
165 | domain_adaptation_images/dslr/images/headphones/frame_0011.jpg 10
166 | domain_adaptation_images/dslr/images/headphones/frame_0012.jpg 10
167 | domain_adaptation_images/dslr/images/headphones/frame_0013.jpg 10
168 | domain_adaptation_images/dslr/images/keyboard/frame_0001.jpg 11
169 | domain_adaptation_images/dslr/images/keyboard/frame_0002.jpg 11
170 | domain_adaptation_images/dslr/images/keyboard/frame_0003.jpg 11
171 | domain_adaptation_images/dslr/images/keyboard/frame_0004.jpg 11
172 | domain_adaptation_images/dslr/images/keyboard/frame_0005.jpg 11
173 | domain_adaptation_images/dslr/images/keyboard/frame_0006.jpg 11
174 | domain_adaptation_images/dslr/images/keyboard/frame_0007.jpg 11
175 | domain_adaptation_images/dslr/images/keyboard/frame_0008.jpg 11
176 | domain_adaptation_images/dslr/images/keyboard/frame_0009.jpg 11
177 | domain_adaptation_images/dslr/images/keyboard/frame_0010.jpg 11
178 | domain_adaptation_images/dslr/images/laptop_computer/frame_0001.jpg 12
179 | domain_adaptation_images/dslr/images/laptop_computer/frame_0002.jpg 12
180 | domain_adaptation_images/dslr/images/laptop_computer/frame_0003.jpg 12
181 | domain_adaptation_images/dslr/images/laptop_computer/frame_0004.jpg 12
182 | domain_adaptation_images/dslr/images/laptop_computer/frame_0005.jpg 12
183 | domain_adaptation_images/dslr/images/laptop_computer/frame_0006.jpg 12
184 | domain_adaptation_images/dslr/images/laptop_computer/frame_0007.jpg 12
185 | domain_adaptation_images/dslr/images/laptop_computer/frame_0008.jpg 12
186 | domain_adaptation_images/dslr/images/laptop_computer/frame_0009.jpg 12
187 | domain_adaptation_images/dslr/images/laptop_computer/frame_0010.jpg 12
188 | domain_adaptation_images/dslr/images/laptop_computer/frame_0011.jpg 12
189 | domain_adaptation_images/dslr/images/laptop_computer/frame_0012.jpg 12
190 | domain_adaptation_images/dslr/images/laptop_computer/frame_0013.jpg 12
191 | domain_adaptation_images/dslr/images/laptop_computer/frame_0014.jpg 12
192 | domain_adaptation_images/dslr/images/laptop_computer/frame_0015.jpg 12
193 | domain_adaptation_images/dslr/images/laptop_computer/frame_0016.jpg 12
194 | domain_adaptation_images/dslr/images/laptop_computer/frame_0017.jpg 12
195 | domain_adaptation_images/dslr/images/laptop_computer/frame_0018.jpg 12
196 | domain_adaptation_images/dslr/images/laptop_computer/frame_0019.jpg 12
197 | domain_adaptation_images/dslr/images/laptop_computer/frame_0020.jpg 12
198 | domain_adaptation_images/dslr/images/laptop_computer/frame_0021.jpg 12
199 | domain_adaptation_images/dslr/images/laptop_computer/frame_0022.jpg 12
200 | domain_adaptation_images/dslr/images/laptop_computer/frame_0023.jpg 12
201 | domain_adaptation_images/dslr/images/laptop_computer/frame_0024.jpg 12
202 | domain_adaptation_images/dslr/images/letter_tray/frame_0001.jpg 13
203 | domain_adaptation_images/dslr/images/letter_tray/frame_0002.jpg 13
204 | domain_adaptation_images/dslr/images/letter_tray/frame_0003.jpg 13
205 | domain_adaptation_images/dslr/images/letter_tray/frame_0004.jpg 13
206 | domain_adaptation_images/dslr/images/letter_tray/frame_0005.jpg 13
207 | domain_adaptation_images/dslr/images/letter_tray/frame_0006.jpg 13
208 | domain_adaptation_images/dslr/images/letter_tray/frame_0007.jpg 13
209 | domain_adaptation_images/dslr/images/letter_tray/frame_0008.jpg 13
210 | domain_adaptation_images/dslr/images/letter_tray/frame_0009.jpg 13
211 | domain_adaptation_images/dslr/images/letter_tray/frame_0010.jpg 13
212 | domain_adaptation_images/dslr/images/letter_tray/frame_0011.jpg 13
213 | domain_adaptation_images/dslr/images/letter_tray/frame_0012.jpg 13
214 | domain_adaptation_images/dslr/images/letter_tray/frame_0013.jpg 13
215 | domain_adaptation_images/dslr/images/letter_tray/frame_0014.jpg 13
216 | domain_adaptation_images/dslr/images/letter_tray/frame_0015.jpg 13
217 | domain_adaptation_images/dslr/images/letter_tray/frame_0016.jpg 13
218 | domain_adaptation_images/dslr/images/mobile_phone/frame_0001.jpg 14
219 | domain_adaptation_images/dslr/images/mobile_phone/frame_0002.jpg 14
220 | domain_adaptation_images/dslr/images/mobile_phone/frame_0003.jpg 14
221 | domain_adaptation_images/dslr/images/mobile_phone/frame_0004.jpg 14
222 | domain_adaptation_images/dslr/images/mobile_phone/frame_0005.jpg 14
223 | domain_adaptation_images/dslr/images/mobile_phone/frame_0006.jpg 14
224 | domain_adaptation_images/dslr/images/mobile_phone/frame_0007.jpg 14
225 | domain_adaptation_images/dslr/images/mobile_phone/frame_0008.jpg 14
226 | domain_adaptation_images/dslr/images/mobile_phone/frame_0009.jpg 14
227 | domain_adaptation_images/dslr/images/mobile_phone/frame_0010.jpg 14
228 | domain_adaptation_images/dslr/images/mobile_phone/frame_0011.jpg 14
229 | domain_adaptation_images/dslr/images/mobile_phone/frame_0012.jpg 14
230 | domain_adaptation_images/dslr/images/mobile_phone/frame_0013.jpg 14
231 | domain_adaptation_images/dslr/images/mobile_phone/frame_0014.jpg 14
232 | domain_adaptation_images/dslr/images/mobile_phone/frame_0015.jpg 14
233 | domain_adaptation_images/dslr/images/mobile_phone/frame_0016.jpg 14
234 | domain_adaptation_images/dslr/images/mobile_phone/frame_0017.jpg 14
235 | domain_adaptation_images/dslr/images/mobile_phone/frame_0018.jpg 14
236 | domain_adaptation_images/dslr/images/mobile_phone/frame_0019.jpg 14
237 | domain_adaptation_images/dslr/images/mobile_phone/frame_0020.jpg 14
238 | domain_adaptation_images/dslr/images/mobile_phone/frame_0021.jpg 14
239 | domain_adaptation_images/dslr/images/mobile_phone/frame_0022.jpg 14
240 | domain_adaptation_images/dslr/images/mobile_phone/frame_0023.jpg 14
241 | domain_adaptation_images/dslr/images/mobile_phone/frame_0024.jpg 14
242 | domain_adaptation_images/dslr/images/mobile_phone/frame_0025.jpg 14
243 | domain_adaptation_images/dslr/images/mobile_phone/frame_0026.jpg 14
244 | domain_adaptation_images/dslr/images/mobile_phone/frame_0027.jpg 14
245 | domain_adaptation_images/dslr/images/mobile_phone/frame_0028.jpg 14
246 | domain_adaptation_images/dslr/images/mobile_phone/frame_0029.jpg 14
247 | domain_adaptation_images/dslr/images/mobile_phone/frame_0030.jpg 14
248 | domain_adaptation_images/dslr/images/mobile_phone/frame_0031.jpg 14
249 | domain_adaptation_images/dslr/images/monitor/frame_0001.jpg 15
250 | domain_adaptation_images/dslr/images/monitor/frame_0002.jpg 15
251 | domain_adaptation_images/dslr/images/monitor/frame_0003.jpg 15
252 | domain_adaptation_images/dslr/images/monitor/frame_0004.jpg 15
253 | domain_adaptation_images/dslr/images/monitor/frame_0005.jpg 15
254 | domain_adaptation_images/dslr/images/monitor/frame_0006.jpg 15
255 | domain_adaptation_images/dslr/images/monitor/frame_0007.jpg 15
256 | domain_adaptation_images/dslr/images/monitor/frame_0008.jpg 15
257 | domain_adaptation_images/dslr/images/monitor/frame_0009.jpg 15
258 | domain_adaptation_images/dslr/images/monitor/frame_0010.jpg 15
259 | domain_adaptation_images/dslr/images/monitor/frame_0011.jpg 15
260 | domain_adaptation_images/dslr/images/monitor/frame_0012.jpg 15
261 | domain_adaptation_images/dslr/images/monitor/frame_0013.jpg 15
262 | domain_adaptation_images/dslr/images/monitor/frame_0014.jpg 15
263 | domain_adaptation_images/dslr/images/monitor/frame_0015.jpg 15
264 | domain_adaptation_images/dslr/images/monitor/frame_0016.jpg 15
265 | domain_adaptation_images/dslr/images/monitor/frame_0017.jpg 15
266 | domain_adaptation_images/dslr/images/monitor/frame_0018.jpg 15
267 | domain_adaptation_images/dslr/images/monitor/frame_0019.jpg 15
268 | domain_adaptation_images/dslr/images/monitor/frame_0020.jpg 15
269 | domain_adaptation_images/dslr/images/monitor/frame_0021.jpg 15
270 | domain_adaptation_images/dslr/images/monitor/frame_0022.jpg 15
271 | domain_adaptation_images/dslr/images/mouse/frame_0001.jpg 16
272 | domain_adaptation_images/dslr/images/mouse/frame_0002.jpg 16
273 | domain_adaptation_images/dslr/images/mouse/frame_0003.jpg 16
274 | domain_adaptation_images/dslr/images/mouse/frame_0004.jpg 16
275 | domain_adaptation_images/dslr/images/mouse/frame_0005.jpg 16
276 | domain_adaptation_images/dslr/images/mouse/frame_0006.jpg 16
277 | domain_adaptation_images/dslr/images/mouse/frame_0007.jpg 16
278 | domain_adaptation_images/dslr/images/mouse/frame_0008.jpg 16
279 | domain_adaptation_images/dslr/images/mouse/frame_0009.jpg 16
280 | domain_adaptation_images/dslr/images/mouse/frame_0010.jpg 16
281 | domain_adaptation_images/dslr/images/mouse/frame_0011.jpg 16
282 | domain_adaptation_images/dslr/images/mouse/frame_0012.jpg 16
283 | domain_adaptation_images/dslr/images/mug/frame_0001.jpg 17
284 | domain_adaptation_images/dslr/images/mug/frame_0002.jpg 17
285 | domain_adaptation_images/dslr/images/mug/frame_0003.jpg 17
286 | domain_adaptation_images/dslr/images/mug/frame_0004.jpg 17
287 | domain_adaptation_images/dslr/images/mug/frame_0005.jpg 17
288 | domain_adaptation_images/dslr/images/mug/frame_0006.jpg 17
289 | domain_adaptation_images/dslr/images/mug/frame_0007.jpg 17
290 | domain_adaptation_images/dslr/images/mug/frame_0008.jpg 17
291 | domain_adaptation_images/dslr/images/paper_notebook/frame_0001.jpg 18
292 | domain_adaptation_images/dslr/images/paper_notebook/frame_0002.jpg 18
293 | domain_adaptation_images/dslr/images/paper_notebook/frame_0003.jpg 18
294 | domain_adaptation_images/dslr/images/paper_notebook/frame_0004.jpg 18
295 | domain_adaptation_images/dslr/images/paper_notebook/frame_0005.jpg 18
296 | domain_adaptation_images/dslr/images/paper_notebook/frame_0006.jpg 18
297 | domain_adaptation_images/dslr/images/paper_notebook/frame_0007.jpg 18
298 | domain_adaptation_images/dslr/images/paper_notebook/frame_0008.jpg 18
299 | domain_adaptation_images/dslr/images/paper_notebook/frame_0009.jpg 18
300 | domain_adaptation_images/dslr/images/paper_notebook/frame_0010.jpg 18
301 | domain_adaptation_images/dslr/images/pen/frame_0001.jpg 19
302 | domain_adaptation_images/dslr/images/pen/frame_0002.jpg 19
303 | domain_adaptation_images/dslr/images/pen/frame_0003.jpg 19
304 | domain_adaptation_images/dslr/images/pen/frame_0004.jpg 19
305 | domain_adaptation_images/dslr/images/pen/frame_0005.jpg 19
306 | domain_adaptation_images/dslr/images/pen/frame_0006.jpg 19
307 | domain_adaptation_images/dslr/images/pen/frame_0007.jpg 19
308 | domain_adaptation_images/dslr/images/pen/frame_0008.jpg 19
309 | domain_adaptation_images/dslr/images/pen/frame_0009.jpg 19
310 | domain_adaptation_images/dslr/images/pen/frame_0010.jpg 19
311 | domain_adaptation_images/dslr/images/phone/frame_0001.jpg 20
312 | domain_adaptation_images/dslr/images/phone/frame_0002.jpg 20
313 | domain_adaptation_images/dslr/images/phone/frame_0003.jpg 20
314 | domain_adaptation_images/dslr/images/phone/frame_0004.jpg 20
315 | domain_adaptation_images/dslr/images/phone/frame_0005.jpg 20
316 | domain_adaptation_images/dslr/images/phone/frame_0006.jpg 20
317 | domain_adaptation_images/dslr/images/phone/frame_0007.jpg 20
318 | domain_adaptation_images/dslr/images/phone/frame_0008.jpg 20
319 | domain_adaptation_images/dslr/images/phone/frame_0009.jpg 20
320 | domain_adaptation_images/dslr/images/phone/frame_0010.jpg 20
321 | domain_adaptation_images/dslr/images/phone/frame_0011.jpg 20
322 | domain_adaptation_images/dslr/images/phone/frame_0012.jpg 20
323 | domain_adaptation_images/dslr/images/phone/frame_0013.jpg 20
324 | domain_adaptation_images/dslr/images/printer/frame_0001.jpg 21
325 | domain_adaptation_images/dslr/images/printer/frame_0002.jpg 21
326 | domain_adaptation_images/dslr/images/printer/frame_0003.jpg 21
327 | domain_adaptation_images/dslr/images/printer/frame_0004.jpg 21
328 | domain_adaptation_images/dslr/images/printer/frame_0005.jpg 21
329 | domain_adaptation_images/dslr/images/printer/frame_0006.jpg 21
330 | domain_adaptation_images/dslr/images/printer/frame_0007.jpg 21
331 | domain_adaptation_images/dslr/images/printer/frame_0008.jpg 21
332 | domain_adaptation_images/dslr/images/printer/frame_0009.jpg 21
333 | domain_adaptation_images/dslr/images/printer/frame_0010.jpg 21
334 | domain_adaptation_images/dslr/images/printer/frame_0011.jpg 21
335 | domain_adaptation_images/dslr/images/printer/frame_0012.jpg 21
336 | domain_adaptation_images/dslr/images/printer/frame_0013.jpg 21
337 | domain_adaptation_images/dslr/images/printer/frame_0014.jpg 21
338 | domain_adaptation_images/dslr/images/printer/frame_0015.jpg 21
339 | domain_adaptation_images/dslr/images/projector/frame_0001.jpg 22
340 | domain_adaptation_images/dslr/images/projector/frame_0002.jpg 22
341 | domain_adaptation_images/dslr/images/projector/frame_0003.jpg 22
342 | domain_adaptation_images/dslr/images/projector/frame_0004.jpg 22
343 | domain_adaptation_images/dslr/images/projector/frame_0005.jpg 22
344 | domain_adaptation_images/dslr/images/projector/frame_0006.jpg 22
345 | domain_adaptation_images/dslr/images/projector/frame_0007.jpg 22
346 | domain_adaptation_images/dslr/images/projector/frame_0008.jpg 22
347 | domain_adaptation_images/dslr/images/projector/frame_0009.jpg 22
348 | domain_adaptation_images/dslr/images/projector/frame_0010.jpg 22
349 | domain_adaptation_images/dslr/images/projector/frame_0011.jpg 22
350 | domain_adaptation_images/dslr/images/projector/frame_0012.jpg 22
351 | domain_adaptation_images/dslr/images/projector/frame_0013.jpg 22
352 | domain_adaptation_images/dslr/images/projector/frame_0014.jpg 22
353 | domain_adaptation_images/dslr/images/projector/frame_0015.jpg 22
354 | domain_adaptation_images/dslr/images/projector/frame_0016.jpg 22
355 | domain_adaptation_images/dslr/images/projector/frame_0017.jpg 22
356 | domain_adaptation_images/dslr/images/projector/frame_0018.jpg 22
357 | domain_adaptation_images/dslr/images/projector/frame_0019.jpg 22
358 | domain_adaptation_images/dslr/images/projector/frame_0020.jpg 22
359 | domain_adaptation_images/dslr/images/projector/frame_0021.jpg 22
360 | domain_adaptation_images/dslr/images/projector/frame_0022.jpg 22
361 | domain_adaptation_images/dslr/images/projector/frame_0023.jpg 22
362 | domain_adaptation_images/dslr/images/punchers/frame_0001.jpg 23
363 | domain_adaptation_images/dslr/images/punchers/frame_0002.jpg 23
364 | domain_adaptation_images/dslr/images/punchers/frame_0003.jpg 23
365 | domain_adaptation_images/dslr/images/punchers/frame_0004.jpg 23
366 | domain_adaptation_images/dslr/images/punchers/frame_0005.jpg 23
367 | domain_adaptation_images/dslr/images/punchers/frame_0006.jpg 23
368 | domain_adaptation_images/dslr/images/punchers/frame_0007.jpg 23
369 | domain_adaptation_images/dslr/images/punchers/frame_0008.jpg 23
370 | domain_adaptation_images/dslr/images/punchers/frame_0009.jpg 23
371 | domain_adaptation_images/dslr/images/punchers/frame_0010.jpg 23
372 | domain_adaptation_images/dslr/images/punchers/frame_0011.jpg 23
373 | domain_adaptation_images/dslr/images/punchers/frame_0012.jpg 23
374 | domain_adaptation_images/dslr/images/punchers/frame_0013.jpg 23
375 | domain_adaptation_images/dslr/images/punchers/frame_0014.jpg 23
376 | domain_adaptation_images/dslr/images/punchers/frame_0015.jpg 23
377 | domain_adaptation_images/dslr/images/punchers/frame_0016.jpg 23
378 | domain_adaptation_images/dslr/images/punchers/frame_0017.jpg 23
379 | domain_adaptation_images/dslr/images/punchers/frame_0018.jpg 23
380 | domain_adaptation_images/dslr/images/ring_binder/frame_0001.jpg 24
381 | domain_adaptation_images/dslr/images/ring_binder/frame_0002.jpg 24
382 | domain_adaptation_images/dslr/images/ring_binder/frame_0003.jpg 24
383 | domain_adaptation_images/dslr/images/ring_binder/frame_0004.jpg 24
384 | domain_adaptation_images/dslr/images/ring_binder/frame_0005.jpg 24
385 | domain_adaptation_images/dslr/images/ring_binder/frame_0006.jpg 24
386 | domain_adaptation_images/dslr/images/ring_binder/frame_0007.jpg 24
387 | domain_adaptation_images/dslr/images/ring_binder/frame_0008.jpg 24
388 | domain_adaptation_images/dslr/images/ring_binder/frame_0009.jpg 24
389 | domain_adaptation_images/dslr/images/ring_binder/frame_0010.jpg 24
390 | domain_adaptation_images/dslr/images/ruler/frame_0001.jpg 25
391 | domain_adaptation_images/dslr/images/ruler/frame_0002.jpg 25
392 | domain_adaptation_images/dslr/images/ruler/frame_0003.jpg 25
393 | domain_adaptation_images/dslr/images/ruler/frame_0004.jpg 25
394 | domain_adaptation_images/dslr/images/ruler/frame_0005.jpg 25
395 | domain_adaptation_images/dslr/images/ruler/frame_0006.jpg 25
396 | domain_adaptation_images/dslr/images/ruler/frame_0007.jpg 25
397 | domain_adaptation_images/dslr/images/scissors/frame_0001.jpg 26
398 | domain_adaptation_images/dslr/images/scissors/frame_0002.jpg 26
399 | domain_adaptation_images/dslr/images/scissors/frame_0003.jpg 26
400 | domain_adaptation_images/dslr/images/scissors/frame_0004.jpg 26
401 | domain_adaptation_images/dslr/images/scissors/frame_0005.jpg 26
402 | domain_adaptation_images/dslr/images/scissors/frame_0006.jpg 26
403 | domain_adaptation_images/dslr/images/scissors/frame_0007.jpg 26
404 | domain_adaptation_images/dslr/images/scissors/frame_0008.jpg 26
405 | domain_adaptation_images/dslr/images/scissors/frame_0009.jpg 26
406 | domain_adaptation_images/dslr/images/scissors/frame_0010.jpg 26
407 | domain_adaptation_images/dslr/images/scissors/frame_0011.jpg 26
408 | domain_adaptation_images/dslr/images/scissors/frame_0012.jpg 26
409 | domain_adaptation_images/dslr/images/scissors/frame_0013.jpg 26
410 | domain_adaptation_images/dslr/images/scissors/frame_0014.jpg 26
411 | domain_adaptation_images/dslr/images/scissors/frame_0015.jpg 26
412 | domain_adaptation_images/dslr/images/scissors/frame_0016.jpg 26
413 | domain_adaptation_images/dslr/images/scissors/frame_0017.jpg 26
414 | domain_adaptation_images/dslr/images/scissors/frame_0018.jpg 26
415 | domain_adaptation_images/dslr/images/speaker/frame_0001.jpg 27
416 | domain_adaptation_images/dslr/images/speaker/frame_0002.jpg 27
417 | domain_adaptation_images/dslr/images/speaker/frame_0003.jpg 27
418 | domain_adaptation_images/dslr/images/speaker/frame_0004.jpg 27
419 | domain_adaptation_images/dslr/images/speaker/frame_0005.jpg 27
420 | domain_adaptation_images/dslr/images/speaker/frame_0006.jpg 27
421 | domain_adaptation_images/dslr/images/speaker/frame_0007.jpg 27
422 | domain_adaptation_images/dslr/images/speaker/frame_0008.jpg 27
423 | domain_adaptation_images/dslr/images/speaker/frame_0009.jpg 27
424 | domain_adaptation_images/dslr/images/speaker/frame_0010.jpg 27
425 | domain_adaptation_images/dslr/images/speaker/frame_0011.jpg 27
426 | domain_adaptation_images/dslr/images/speaker/frame_0012.jpg 27
427 | domain_adaptation_images/dslr/images/speaker/frame_0013.jpg 27
428 | domain_adaptation_images/dslr/images/speaker/frame_0014.jpg 27
429 | domain_adaptation_images/dslr/images/speaker/frame_0015.jpg 27
430 | domain_adaptation_images/dslr/images/speaker/frame_0016.jpg 27
431 | domain_adaptation_images/dslr/images/speaker/frame_0017.jpg 27
432 | domain_adaptation_images/dslr/images/speaker/frame_0018.jpg 27
433 | domain_adaptation_images/dslr/images/speaker/frame_0019.jpg 27
434 | domain_adaptation_images/dslr/images/speaker/frame_0020.jpg 27
435 | domain_adaptation_images/dslr/images/speaker/frame_0021.jpg 27
436 | domain_adaptation_images/dslr/images/speaker/frame_0022.jpg 27
437 | domain_adaptation_images/dslr/images/speaker/frame_0023.jpg 27
438 | domain_adaptation_images/dslr/images/speaker/frame_0024.jpg 27
439 | domain_adaptation_images/dslr/images/speaker/frame_0025.jpg 27
440 | domain_adaptation_images/dslr/images/speaker/frame_0026.jpg 27
441 | domain_adaptation_images/dslr/images/stapler/frame_0001.jpg 28
442 | domain_adaptation_images/dslr/images/stapler/frame_0002.jpg 28
443 | domain_adaptation_images/dslr/images/stapler/frame_0003.jpg 28
444 | domain_adaptation_images/dslr/images/stapler/frame_0004.jpg 28
445 | domain_adaptation_images/dslr/images/stapler/frame_0005.jpg 28
446 | domain_adaptation_images/dslr/images/stapler/frame_0006.jpg 28
447 | domain_adaptation_images/dslr/images/stapler/frame_0007.jpg 28
448 | domain_adaptation_images/dslr/images/stapler/frame_0008.jpg 28
449 | domain_adaptation_images/dslr/images/stapler/frame_0009.jpg 28
450 | domain_adaptation_images/dslr/images/stapler/frame_0010.jpg 28
451 | domain_adaptation_images/dslr/images/stapler/frame_0011.jpg 28
452 | domain_adaptation_images/dslr/images/stapler/frame_0012.jpg 28
453 | domain_adaptation_images/dslr/images/stapler/frame_0013.jpg 28
454 | domain_adaptation_images/dslr/images/stapler/frame_0014.jpg 28
455 | domain_adaptation_images/dslr/images/stapler/frame_0015.jpg 28
456 | domain_adaptation_images/dslr/images/stapler/frame_0016.jpg 28
457 | domain_adaptation_images/dslr/images/stapler/frame_0017.jpg 28
458 | domain_adaptation_images/dslr/images/stapler/frame_0018.jpg 28
459 | domain_adaptation_images/dslr/images/stapler/frame_0019.jpg 28
460 | domain_adaptation_images/dslr/images/stapler/frame_0020.jpg 28
461 | domain_adaptation_images/dslr/images/stapler/frame_0021.jpg 28
462 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0001.jpg 29
463 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0002.jpg 29
464 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0003.jpg 29
465 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0004.jpg 29
466 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0005.jpg 29
467 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0006.jpg 29
468 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0007.jpg 29
469 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0008.jpg 29
470 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0009.jpg 29
471 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0010.jpg 29
472 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0011.jpg 29
473 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0012.jpg 29
474 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0013.jpg 29
475 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0014.jpg 29
476 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0015.jpg 29
477 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0016.jpg 29
478 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0017.jpg 29
479 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0018.jpg 29
480 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0019.jpg 29
481 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0020.jpg 29
482 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0021.jpg 29
483 | domain_adaptation_images/dslr/images/tape_dispenser/frame_0022.jpg 29
484 | domain_adaptation_images/dslr/images/trash_can/frame_0001.jpg 30
485 | domain_adaptation_images/dslr/images/trash_can/frame_0002.jpg 30
486 | domain_adaptation_images/dslr/images/trash_can/frame_0003.jpg 30
487 | domain_adaptation_images/dslr/images/trash_can/frame_0004.jpg 30
488 | domain_adaptation_images/dslr/images/trash_can/frame_0005.jpg 30
489 | domain_adaptation_images/dslr/images/trash_can/frame_0006.jpg 30
490 | domain_adaptation_images/dslr/images/trash_can/frame_0007.jpg 30
491 | domain_adaptation_images/dslr/images/trash_can/frame_0008.jpg 30
492 | domain_adaptation_images/dslr/images/trash_can/frame_0009.jpg 30
493 | domain_adaptation_images/dslr/images/trash_can/frame_0010.jpg 30
494 | domain_adaptation_images/dslr/images/trash_can/frame_0011.jpg 30
495 | domain_adaptation_images/dslr/images/trash_can/frame_0012.jpg 30
496 | domain_adaptation_images/dslr/images/trash_can/frame_0013.jpg 30
497 | domain_adaptation_images/dslr/images/trash_can/frame_0014.jpg 30
498 | domain_adaptation_images/dslr/images/trash_can/frame_0015.jpg 30
499 | 


--------------------------------------------------------------------------------
/data/office/webcam.txt:
--------------------------------------------------------------------------------
  1 | domain_adaptation_images/webcam/images/back_pack/frame_0001.jpg 0
  2 | domain_adaptation_images/webcam/images/back_pack/frame_0002.jpg 0
  3 | domain_adaptation_images/webcam/images/back_pack/frame_0003.jpg 0
  4 | domain_adaptation_images/webcam/images/back_pack/frame_0004.jpg 0
  5 | domain_adaptation_images/webcam/images/back_pack/frame_0005.jpg 0
  6 | domain_adaptation_images/webcam/images/back_pack/frame_0006.jpg 0
  7 | domain_adaptation_images/webcam/images/back_pack/frame_0007.jpg 0
  8 | domain_adaptation_images/webcam/images/back_pack/frame_0008.jpg 0
  9 | domain_adaptation_images/webcam/images/back_pack/frame_0009.jpg 0
 10 | domain_adaptation_images/webcam/images/back_pack/frame_0010.jpg 0
 11 | domain_adaptation_images/webcam/images/back_pack/frame_0011.jpg 0
 12 | domain_adaptation_images/webcam/images/back_pack/frame_0012.jpg 0
 13 | domain_adaptation_images/webcam/images/back_pack/frame_0013.jpg 0
 14 | domain_adaptation_images/webcam/images/back_pack/frame_0014.jpg 0
 15 | domain_adaptation_images/webcam/images/back_pack/frame_0015.jpg 0
 16 | domain_adaptation_images/webcam/images/back_pack/frame_0016.jpg 0
 17 | domain_adaptation_images/webcam/images/back_pack/frame_0017.jpg 0
 18 | domain_adaptation_images/webcam/images/back_pack/frame_0018.jpg 0
 19 | domain_adaptation_images/webcam/images/back_pack/frame_0019.jpg 0
 20 | domain_adaptation_images/webcam/images/back_pack/frame_0020.jpg 0
 21 | domain_adaptation_images/webcam/images/back_pack/frame_0021.jpg 0
 22 | domain_adaptation_images/webcam/images/back_pack/frame_0022.jpg 0
 23 | domain_adaptation_images/webcam/images/back_pack/frame_0023.jpg 0
 24 | domain_adaptation_images/webcam/images/back_pack/frame_0024.jpg 0
 25 | domain_adaptation_images/webcam/images/back_pack/frame_0025.jpg 0
 26 | domain_adaptation_images/webcam/images/back_pack/frame_0026.jpg 0
 27 | domain_adaptation_images/webcam/images/back_pack/frame_0027.jpg 0
 28 | domain_adaptation_images/webcam/images/back_pack/frame_0028.jpg 0
 29 | domain_adaptation_images/webcam/images/back_pack/frame_0029.jpg 0
 30 | domain_adaptation_images/webcam/images/bike/frame_0001.jpg 1
 31 | domain_adaptation_images/webcam/images/bike/frame_0002.jpg 1
 32 | domain_adaptation_images/webcam/images/bike/frame_0003.jpg 1
 33 | domain_adaptation_images/webcam/images/bike/frame_0004.jpg 1
 34 | domain_adaptation_images/webcam/images/bike/frame_0005.jpg 1
 35 | domain_adaptation_images/webcam/images/bike/frame_0006.jpg 1
 36 | domain_adaptation_images/webcam/images/bike/frame_0007.jpg 1
 37 | domain_adaptation_images/webcam/images/bike/frame_0008.jpg 1
 38 | domain_adaptation_images/webcam/images/bike/frame_0009.jpg 1
 39 | domain_adaptation_images/webcam/images/bike/frame_0010.jpg 1
 40 | domain_adaptation_images/webcam/images/bike/frame_0011.jpg 1
 41 | domain_adaptation_images/webcam/images/bike/frame_0012.jpg 1
 42 | domain_adaptation_images/webcam/images/bike/frame_0013.jpg 1
 43 | domain_adaptation_images/webcam/images/bike/frame_0014.jpg 1
 44 | domain_adaptation_images/webcam/images/bike/frame_0015.jpg 1
 45 | domain_adaptation_images/webcam/images/bike/frame_0016.jpg 1
 46 | domain_adaptation_images/webcam/images/bike/frame_0017.jpg 1
 47 | domain_adaptation_images/webcam/images/bike/frame_0018.jpg 1
 48 | domain_adaptation_images/webcam/images/bike/frame_0019.jpg 1
 49 | domain_adaptation_images/webcam/images/bike/frame_0020.jpg 1
 50 | domain_adaptation_images/webcam/images/bike/frame_0021.jpg 1
 51 | domain_adaptation_images/webcam/images/bike_helmet/frame_0001.jpg 2
 52 | domain_adaptation_images/webcam/images/bike_helmet/frame_0002.jpg 2
 53 | domain_adaptation_images/webcam/images/bike_helmet/frame_0003.jpg 2
 54 | domain_adaptation_images/webcam/images/bike_helmet/frame_0004.jpg 2
 55 | domain_adaptation_images/webcam/images/bike_helmet/frame_0005.jpg 2
 56 | domain_adaptation_images/webcam/images/bike_helmet/frame_0006.jpg 2
 57 | domain_adaptation_images/webcam/images/bike_helmet/frame_0007.jpg 2
 58 | domain_adaptation_images/webcam/images/bike_helmet/frame_0008.jpg 2
 59 | domain_adaptation_images/webcam/images/bike_helmet/frame_0009.jpg 2
 60 | domain_adaptation_images/webcam/images/bike_helmet/frame_0010.jpg 2
 61 | domain_adaptation_images/webcam/images/bike_helmet/frame_0011.jpg 2
 62 | domain_adaptation_images/webcam/images/bike_helmet/frame_0012.jpg 2
 63 | domain_adaptation_images/webcam/images/bike_helmet/frame_0013.jpg 2
 64 | domain_adaptation_images/webcam/images/bike_helmet/frame_0014.jpg 2
 65 | domain_adaptation_images/webcam/images/bike_helmet/frame_0015.jpg 2
 66 | domain_adaptation_images/webcam/images/bike_helmet/frame_0016.jpg 2
 67 | domain_adaptation_images/webcam/images/bike_helmet/frame_0017.jpg 2
 68 | domain_adaptation_images/webcam/images/bike_helmet/frame_0018.jpg 2
 69 | domain_adaptation_images/webcam/images/bike_helmet/frame_0019.jpg 2
 70 | domain_adaptation_images/webcam/images/bike_helmet/frame_0020.jpg 2
 71 | domain_adaptation_images/webcam/images/bike_helmet/frame_0021.jpg 2
 72 | domain_adaptation_images/webcam/images/bike_helmet/frame_0022.jpg 2
 73 | domain_adaptation_images/webcam/images/bike_helmet/frame_0023.jpg 2
 74 | domain_adaptation_images/webcam/images/bike_helmet/frame_0024.jpg 2
 75 | domain_adaptation_images/webcam/images/bike_helmet/frame_0025.jpg 2
 76 | domain_adaptation_images/webcam/images/bike_helmet/frame_0026.jpg 2
 77 | domain_adaptation_images/webcam/images/bike_helmet/frame_0027.jpg 2
 78 | domain_adaptation_images/webcam/images/bike_helmet/frame_0028.jpg 2
 79 | domain_adaptation_images/webcam/images/bookcase/frame_0001.jpg 3
 80 | domain_adaptation_images/webcam/images/bookcase/frame_0002.jpg 3
 81 | domain_adaptation_images/webcam/images/bookcase/frame_0003.jpg 3
 82 | domain_adaptation_images/webcam/images/bookcase/frame_0004.jpg 3
 83 | domain_adaptation_images/webcam/images/bookcase/frame_0005.jpg 3
 84 | domain_adaptation_images/webcam/images/bookcase/frame_0006.jpg 3
 85 | domain_adaptation_images/webcam/images/bookcase/frame_0007.jpg 3
 86 | domain_adaptation_images/webcam/images/bookcase/frame_0008.jpg 3
 87 | domain_adaptation_images/webcam/images/bookcase/frame_0009.jpg 3
 88 | domain_adaptation_images/webcam/images/bookcase/frame_0010.jpg 3
 89 | domain_adaptation_images/webcam/images/bookcase/frame_0011.jpg 3
 90 | domain_adaptation_images/webcam/images/bookcase/frame_0012.jpg 3
 91 | domain_adaptation_images/webcam/images/bottle/frame_0001.jpg 4
 92 | domain_adaptation_images/webcam/images/bottle/frame_0002.jpg 4
 93 | domain_adaptation_images/webcam/images/bottle/frame_0003.jpg 4
 94 | domain_adaptation_images/webcam/images/bottle/frame_0004.jpg 4
 95 | domain_adaptation_images/webcam/images/bottle/frame_0005.jpg 4
 96 | domain_adaptation_images/webcam/images/bottle/frame_0006.jpg 4
 97 | domain_adaptation_images/webcam/images/bottle/frame_0007.jpg 4
 98 | domain_adaptation_images/webcam/images/bottle/frame_0008.jpg 4
 99 | domain_adaptation_images/webcam/images/bottle/frame_0009.jpg 4
100 | domain_adaptation_images/webcam/images/bottle/frame_0010.jpg 4
101 | domain_adaptation_images/webcam/images/bottle/frame_0011.jpg 4
102 | domain_adaptation_images/webcam/images/bottle/frame_0012.jpg 4
103 | domain_adaptation_images/webcam/images/bottle/frame_0013.jpg 4
104 | domain_adaptation_images/webcam/images/bottle/frame_0014.jpg 4
105 | domain_adaptation_images/webcam/images/bottle/frame_0015.jpg 4
106 | domain_adaptation_images/webcam/images/bottle/frame_0016.jpg 4
107 | domain_adaptation_images/webcam/images/calculator/frame_0001.jpg 5
108 | domain_adaptation_images/webcam/images/calculator/frame_0002.jpg 5
109 | domain_adaptation_images/webcam/images/calculator/frame_0003.jpg 5
110 | domain_adaptation_images/webcam/images/calculator/frame_0004.jpg 5
111 | domain_adaptation_images/webcam/images/calculator/frame_0005.jpg 5
112 | domain_adaptation_images/webcam/images/calculator/frame_0006.jpg 5
113 | domain_adaptation_images/webcam/images/calculator/frame_0007.jpg 5
114 | domain_adaptation_images/webcam/images/calculator/frame_0008.jpg 5
115 | domain_adaptation_images/webcam/images/calculator/frame_0009.jpg 5
116 | domain_adaptation_images/webcam/images/calculator/frame_0010.jpg 5
117 | domain_adaptation_images/webcam/images/calculator/frame_0011.jpg 5
118 | domain_adaptation_images/webcam/images/calculator/frame_0012.jpg 5
119 | domain_adaptation_images/webcam/images/calculator/frame_0013.jpg 5
120 | domain_adaptation_images/webcam/images/calculator/frame_0014.jpg 5
121 | domain_adaptation_images/webcam/images/calculator/frame_0015.jpg 5
122 | domain_adaptation_images/webcam/images/calculator/frame_0016.jpg 5
123 | domain_adaptation_images/webcam/images/calculator/frame_0017.jpg 5
124 | domain_adaptation_images/webcam/images/calculator/frame_0018.jpg 5
125 | domain_adaptation_images/webcam/images/calculator/frame_0019.jpg 5
126 | domain_adaptation_images/webcam/images/calculator/frame_0020.jpg 5
127 | domain_adaptation_images/webcam/images/calculator/frame_0021.jpg 5
128 | domain_adaptation_images/webcam/images/calculator/frame_0022.jpg 5
129 | domain_adaptation_images/webcam/images/calculator/frame_0023.jpg 5
130 | domain_adaptation_images/webcam/images/calculator/frame_0024.jpg 5
131 | domain_adaptation_images/webcam/images/calculator/frame_0025.jpg 5
132 | domain_adaptation_images/webcam/images/calculator/frame_0026.jpg 5
133 | domain_adaptation_images/webcam/images/calculator/frame_0027.jpg 5
134 | domain_adaptation_images/webcam/images/calculator/frame_0028.jpg 5
135 | domain_adaptation_images/webcam/images/calculator/frame_0029.jpg 5
136 | domain_adaptation_images/webcam/images/calculator/frame_0030.jpg 5
137 | domain_adaptation_images/webcam/images/calculator/frame_0031.jpg 5
138 | domain_adaptation_images/webcam/images/desk_chair/frame_0001.jpg 6
139 | domain_adaptation_images/webcam/images/desk_chair/frame_0002.jpg 6
140 | domain_adaptation_images/webcam/images/desk_chair/frame_0003.jpg 6
141 | domain_adaptation_images/webcam/images/desk_chair/frame_0004.jpg 6
142 | domain_adaptation_images/webcam/images/desk_chair/frame_0005.jpg 6
143 | domain_adaptation_images/webcam/images/desk_chair/frame_0006.jpg 6
144 | domain_adaptation_images/webcam/images/desk_chair/frame_0007.jpg 6
145 | domain_adaptation_images/webcam/images/desk_chair/frame_0008.jpg 6
146 | domain_adaptation_images/webcam/images/desk_chair/frame_0009.jpg 6
147 | domain_adaptation_images/webcam/images/desk_chair/frame_0010.jpg 6
148 | domain_adaptation_images/webcam/images/desk_chair/frame_0011.jpg 6
149 | domain_adaptation_images/webcam/images/desk_chair/frame_0012.jpg 6
150 | domain_adaptation_images/webcam/images/desk_chair/frame_0013.jpg 6
151 | domain_adaptation_images/webcam/images/desk_chair/frame_0014.jpg 6
152 | domain_adaptation_images/webcam/images/desk_chair/frame_0015.jpg 6
153 | domain_adaptation_images/webcam/images/desk_chair/frame_0016.jpg 6
154 | domain_adaptation_images/webcam/images/desk_chair/frame_0017.jpg 6
155 | domain_adaptation_images/webcam/images/desk_chair/frame_0018.jpg 6
156 | domain_adaptation_images/webcam/images/desk_chair/frame_0019.jpg 6
157 | domain_adaptation_images/webcam/images/desk_chair/frame_0020.jpg 6
158 | domain_adaptation_images/webcam/images/desk_chair/frame_0021.jpg 6
159 | domain_adaptation_images/webcam/images/desk_chair/frame_0022.jpg 6
160 | domain_adaptation_images/webcam/images/desk_chair/frame_0023.jpg 6
161 | domain_adaptation_images/webcam/images/desk_chair/frame_0024.jpg 6
162 | domain_adaptation_images/webcam/images/desk_chair/frame_0025.jpg 6
163 | domain_adaptation_images/webcam/images/desk_chair/frame_0026.jpg 6
164 | domain_adaptation_images/webcam/images/desk_chair/frame_0027.jpg 6
165 | domain_adaptation_images/webcam/images/desk_chair/frame_0028.jpg 6
166 | domain_adaptation_images/webcam/images/desk_chair/frame_0029.jpg 6
167 | domain_adaptation_images/webcam/images/desk_chair/frame_0030.jpg 6
168 | domain_adaptation_images/webcam/images/desk_chair/frame_0031.jpg 6
169 | domain_adaptation_images/webcam/images/desk_chair/frame_0032.jpg 6
170 | domain_adaptation_images/webcam/images/desk_chair/frame_0033.jpg 6
171 | domain_adaptation_images/webcam/images/desk_chair/frame_0034.jpg 6
172 | domain_adaptation_images/webcam/images/desk_chair/frame_0035.jpg 6
173 | domain_adaptation_images/webcam/images/desk_chair/frame_0036.jpg 6
174 | domain_adaptation_images/webcam/images/desk_chair/frame_0037.jpg 6
175 | domain_adaptation_images/webcam/images/desk_chair/frame_0038.jpg 6
176 | domain_adaptation_images/webcam/images/desk_chair/frame_0039.jpg 6
177 | domain_adaptation_images/webcam/images/desk_chair/frame_0040.jpg 6
178 | domain_adaptation_images/webcam/images/desk_lamp/frame_0001.jpg 7
179 | domain_adaptation_images/webcam/images/desk_lamp/frame_0002.jpg 7
180 | domain_adaptation_images/webcam/images/desk_lamp/frame_0003.jpg 7
181 | domain_adaptation_images/webcam/images/desk_lamp/frame_0004.jpg 7
182 | domain_adaptation_images/webcam/images/desk_lamp/frame_0005.jpg 7
183 | domain_adaptation_images/webcam/images/desk_lamp/frame_0006.jpg 7
184 | domain_adaptation_images/webcam/images/desk_lamp/frame_0007.jpg 7
185 | domain_adaptation_images/webcam/images/desk_lamp/frame_0008.jpg 7
186 | domain_adaptation_images/webcam/images/desk_lamp/frame_0009.jpg 7
187 | domain_adaptation_images/webcam/images/desk_lamp/frame_0010.jpg 7
188 | domain_adaptation_images/webcam/images/desk_lamp/frame_0011.jpg 7
189 | domain_adaptation_images/webcam/images/desk_lamp/frame_0012.jpg 7
190 | domain_adaptation_images/webcam/images/desk_lamp/frame_0013.jpg 7
191 | domain_adaptation_images/webcam/images/desk_lamp/frame_0014.jpg 7
192 | domain_adaptation_images/webcam/images/desk_lamp/frame_0015.jpg 7
193 | domain_adaptation_images/webcam/images/desk_lamp/frame_0016.jpg 7
194 | domain_adaptation_images/webcam/images/desk_lamp/frame_0017.jpg 7
195 | domain_adaptation_images/webcam/images/desk_lamp/frame_0018.jpg 7
196 | domain_adaptation_images/webcam/images/desktop_computer/frame_0001.jpg 8
197 | domain_adaptation_images/webcam/images/desktop_computer/frame_0002.jpg 8
198 | domain_adaptation_images/webcam/images/desktop_computer/frame_0003.jpg 8
199 | domain_adaptation_images/webcam/images/desktop_computer/frame_0004.jpg 8
200 | domain_adaptation_images/webcam/images/desktop_computer/frame_0005.jpg 8
201 | domain_adaptation_images/webcam/images/desktop_computer/frame_0006.jpg 8
202 | domain_adaptation_images/webcam/images/desktop_computer/frame_0007.jpg 8
203 | domain_adaptation_images/webcam/images/desktop_computer/frame_0008.jpg 8
204 | domain_adaptation_images/webcam/images/desktop_computer/frame_0009.jpg 8
205 | domain_adaptation_images/webcam/images/desktop_computer/frame_0010.jpg 8
206 | domain_adaptation_images/webcam/images/desktop_computer/frame_0011.jpg 8
207 | domain_adaptation_images/webcam/images/desktop_computer/frame_0012.jpg 8
208 | domain_adaptation_images/webcam/images/desktop_computer/frame_0013.jpg 8
209 | domain_adaptation_images/webcam/images/desktop_computer/frame_0014.jpg 8
210 | domain_adaptation_images/webcam/images/desktop_computer/frame_0015.jpg 8
211 | domain_adaptation_images/webcam/images/desktop_computer/frame_0016.jpg 8
212 | domain_adaptation_images/webcam/images/desktop_computer/frame_0017.jpg 8
213 | domain_adaptation_images/webcam/images/desktop_computer/frame_0018.jpg 8
214 | domain_adaptation_images/webcam/images/desktop_computer/frame_0019.jpg 8
215 | domain_adaptation_images/webcam/images/desktop_computer/frame_0020.jpg 8
216 | domain_adaptation_images/webcam/images/desktop_computer/frame_0021.jpg 8
217 | domain_adaptation_images/webcam/images/file_cabinet/frame_0001.jpg 9
218 | domain_adaptation_images/webcam/images/file_cabinet/frame_0002.jpg 9
219 | domain_adaptation_images/webcam/images/file_cabinet/frame_0003.jpg 9
220 | domain_adaptation_images/webcam/images/file_cabinet/frame_0004.jpg 9
221 | domain_adaptation_images/webcam/images/file_cabinet/frame_0005.jpg 9
222 | domain_adaptation_images/webcam/images/file_cabinet/frame_0006.jpg 9
223 | domain_adaptation_images/webcam/images/file_cabinet/frame_0007.jpg 9
224 | domain_adaptation_images/webcam/images/file_cabinet/frame_0008.jpg 9
225 | domain_adaptation_images/webcam/images/file_cabinet/frame_0009.jpg 9
226 | domain_adaptation_images/webcam/images/file_cabinet/frame_0010.jpg 9
227 | domain_adaptation_images/webcam/images/file_cabinet/frame_0011.jpg 9
228 | domain_adaptation_images/webcam/images/file_cabinet/frame_0012.jpg 9
229 | domain_adaptation_images/webcam/images/file_cabinet/frame_0013.jpg 9
230 | domain_adaptation_images/webcam/images/file_cabinet/frame_0014.jpg 9
231 | domain_adaptation_images/webcam/images/file_cabinet/frame_0015.jpg 9
232 | domain_adaptation_images/webcam/images/file_cabinet/frame_0016.jpg 9
233 | domain_adaptation_images/webcam/images/file_cabinet/frame_0017.jpg 9
234 | domain_adaptation_images/webcam/images/file_cabinet/frame_0018.jpg 9
235 | domain_adaptation_images/webcam/images/file_cabinet/frame_0019.jpg 9
236 | domain_adaptation_images/webcam/images/headphones/frame_0001.jpg 10
237 | domain_adaptation_images/webcam/images/headphones/frame_0002.jpg 10
238 | domain_adaptation_images/webcam/images/headphones/frame_0003.jpg 10
239 | domain_adaptation_images/webcam/images/headphones/frame_0004.jpg 10
240 | domain_adaptation_images/webcam/images/headphones/frame_0005.jpg 10
241 | domain_adaptation_images/webcam/images/headphones/frame_0006.jpg 10
242 | domain_adaptation_images/webcam/images/headphones/frame_0007.jpg 10
243 | domain_adaptation_images/webcam/images/headphones/frame_0008.jpg 10
244 | domain_adaptation_images/webcam/images/headphones/frame_0009.jpg 10
245 | domain_adaptation_images/webcam/images/headphones/frame_0010.jpg 10
246 | domain_adaptation_images/webcam/images/headphones/frame_0011.jpg 10
247 | domain_adaptation_images/webcam/images/headphones/frame_0012.jpg 10
248 | domain_adaptation_images/webcam/images/headphones/frame_0013.jpg 10
249 | domain_adaptation_images/webcam/images/headphones/frame_0014.jpg 10
250 | domain_adaptation_images/webcam/images/headphones/frame_0015.jpg 10
251 | domain_adaptation_images/webcam/images/headphones/frame_0016.jpg 10
252 | domain_adaptation_images/webcam/images/headphones/frame_0017.jpg 10
253 | domain_adaptation_images/webcam/images/headphones/frame_0018.jpg 10
254 | domain_adaptation_images/webcam/images/headphones/frame_0019.jpg 10
255 | domain_adaptation_images/webcam/images/headphones/frame_0020.jpg 10
256 | domain_adaptation_images/webcam/images/headphones/frame_0021.jpg 10
257 | domain_adaptation_images/webcam/images/headphones/frame_0022.jpg 10
258 | domain_adaptation_images/webcam/images/headphones/frame_0023.jpg 10
259 | domain_adaptation_images/webcam/images/headphones/frame_0024.jpg 10
260 | domain_adaptation_images/webcam/images/headphones/frame_0025.jpg 10
261 | domain_adaptation_images/webcam/images/headphones/frame_0026.jpg 10
262 | domain_adaptation_images/webcam/images/headphones/frame_0027.jpg 10
263 | domain_adaptation_images/webcam/images/keyboard/frame_0001.jpg 11
264 | domain_adaptation_images/webcam/images/keyboard/frame_0002.jpg 11
265 | domain_adaptation_images/webcam/images/keyboard/frame_0003.jpg 11
266 | domain_adaptation_images/webcam/images/keyboard/frame_0004.jpg 11
267 | domain_adaptation_images/webcam/images/keyboard/frame_0005.jpg 11
268 | domain_adaptation_images/webcam/images/keyboard/frame_0006.jpg 11
269 | domain_adaptation_images/webcam/images/keyboard/frame_0007.jpg 11
270 | domain_adaptation_images/webcam/images/keyboard/frame_0008.jpg 11
271 | domain_adaptation_images/webcam/images/keyboard/frame_0009.jpg 11
272 | domain_adaptation_images/webcam/images/keyboard/frame_0010.jpg 11
273 | domain_adaptation_images/webcam/images/keyboard/frame_0011.jpg 11
274 | domain_adaptation_images/webcam/images/keyboard/frame_0012.jpg 11
275 | domain_adaptation_images/webcam/images/keyboard/frame_0013.jpg 11
276 | domain_adaptation_images/webcam/images/keyboard/frame_0014.jpg 11
277 | domain_adaptation_images/webcam/images/keyboard/frame_0015.jpg 11
278 | domain_adaptation_images/webcam/images/keyboard/frame_0016.jpg 11
279 | domain_adaptation_images/webcam/images/keyboard/frame_0017.jpg 11
280 | domain_adaptation_images/webcam/images/keyboard/frame_0018.jpg 11
281 | domain_adaptation_images/webcam/images/keyboard/frame_0019.jpg 11
282 | domain_adaptation_images/webcam/images/keyboard/frame_0020.jpg 11
283 | domain_adaptation_images/webcam/images/keyboard/frame_0021.jpg 11
284 | domain_adaptation_images/webcam/images/keyboard/frame_0022.jpg 11
285 | domain_adaptation_images/webcam/images/keyboard/frame_0023.jpg 11
286 | domain_adaptation_images/webcam/images/keyboard/frame_0024.jpg 11
287 | domain_adaptation_images/webcam/images/keyboard/frame_0025.jpg 11
288 | domain_adaptation_images/webcam/images/keyboard/frame_0026.jpg 11
289 | domain_adaptation_images/webcam/images/keyboard/frame_0027.jpg 11
290 | domain_adaptation_images/webcam/images/laptop_computer/frame_0001.jpg 12
291 | domain_adaptation_images/webcam/images/laptop_computer/frame_0002.jpg 12
292 | domain_adaptation_images/webcam/images/laptop_computer/frame_0003.jpg 12
293 | domain_adaptation_images/webcam/images/laptop_computer/frame_0004.jpg 12
294 | domain_adaptation_images/webcam/images/laptop_computer/frame_0005.jpg 12
295 | domain_adaptation_images/webcam/images/laptop_computer/frame_0006.jpg 12
296 | domain_adaptation_images/webcam/images/laptop_computer/frame_0007.jpg 12
297 | domain_adaptation_images/webcam/images/laptop_computer/frame_0008.jpg 12
298 | domain_adaptation_images/webcam/images/laptop_computer/frame_0009.jpg 12
299 | domain_adaptation_images/webcam/images/laptop_computer/frame_0010.jpg 12
300 | domain_adaptation_images/webcam/images/laptop_computer/frame_0011.jpg 12
301 | domain_adaptation_images/webcam/images/laptop_computer/frame_0012.jpg 12
302 | domain_adaptation_images/webcam/images/laptop_computer/frame_0013.jpg 12
303 | domain_adaptation_images/webcam/images/laptop_computer/frame_0014.jpg 12
304 | domain_adaptation_images/webcam/images/laptop_computer/frame_0015.jpg 12
305 | domain_adaptation_images/webcam/images/laptop_computer/frame_0016.jpg 12
306 | domain_adaptation_images/webcam/images/laptop_computer/frame_0017.jpg 12
307 | domain_adaptation_images/webcam/images/laptop_computer/frame_0018.jpg 12
308 | domain_adaptation_images/webcam/images/laptop_computer/frame_0019.jpg 12
309 | domain_adaptation_images/webcam/images/laptop_computer/frame_0020.jpg 12
310 | domain_adaptation_images/webcam/images/laptop_computer/frame_0021.jpg 12
311 | domain_adaptation_images/webcam/images/laptop_computer/frame_0022.jpg 12
312 | domain_adaptation_images/webcam/images/laptop_computer/frame_0023.jpg 12
313 | domain_adaptation_images/webcam/images/laptop_computer/frame_0024.jpg 12
314 | domain_adaptation_images/webcam/images/laptop_computer/frame_0025.jpg 12
315 | domain_adaptation_images/webcam/images/laptop_computer/frame_0026.jpg 12
316 | domain_adaptation_images/webcam/images/laptop_computer/frame_0027.jpg 12
317 | domain_adaptation_images/webcam/images/laptop_computer/frame_0028.jpg 12
318 | domain_adaptation_images/webcam/images/laptop_computer/frame_0029.jpg 12
319 | domain_adaptation_images/webcam/images/laptop_computer/frame_0030.jpg 12
320 | domain_adaptation_images/webcam/images/letter_tray/frame_0001.jpg 13
321 | domain_adaptation_images/webcam/images/letter_tray/frame_0002.jpg 13
322 | domain_adaptation_images/webcam/images/letter_tray/frame_0003.jpg 13
323 | domain_adaptation_images/webcam/images/letter_tray/frame_0004.jpg 13
324 | domain_adaptation_images/webcam/images/letter_tray/frame_0005.jpg 13
325 | domain_adaptation_images/webcam/images/letter_tray/frame_0006.jpg 13
326 | domain_adaptation_images/webcam/images/letter_tray/frame_0007.jpg 13
327 | domain_adaptation_images/webcam/images/letter_tray/frame_0008.jpg 13
328 | domain_adaptation_images/webcam/images/letter_tray/frame_0009.jpg 13
329 | domain_adaptation_images/webcam/images/letter_tray/frame_0010.jpg 13
330 | domain_adaptation_images/webcam/images/letter_tray/frame_0011.jpg 13
331 | domain_adaptation_images/webcam/images/letter_tray/frame_0012.jpg 13
332 | domain_adaptation_images/webcam/images/letter_tray/frame_0013.jpg 13
333 | domain_adaptation_images/webcam/images/letter_tray/frame_0014.jpg 13
334 | domain_adaptation_images/webcam/images/letter_tray/frame_0015.jpg 13
335 | domain_adaptation_images/webcam/images/letter_tray/frame_0016.jpg 13
336 | domain_adaptation_images/webcam/images/letter_tray/frame_0017.jpg 13
337 | domain_adaptation_images/webcam/images/letter_tray/frame_0018.jpg 13
338 | domain_adaptation_images/webcam/images/letter_tray/frame_0019.jpg 13
339 | domain_adaptation_images/webcam/images/mobile_phone/frame_0001.jpg 14
340 | domain_adaptation_images/webcam/images/mobile_phone/frame_0002.jpg 14
341 | domain_adaptation_images/webcam/images/mobile_phone/frame_0003.jpg 14
342 | domain_adaptation_images/webcam/images/mobile_phone/frame_0004.jpg 14
343 | domain_adaptation_images/webcam/images/mobile_phone/frame_0005.jpg 14
344 | domain_adaptation_images/webcam/images/mobile_phone/frame_0006.jpg 14
345 | domain_adaptation_images/webcam/images/mobile_phone/frame_0007.jpg 14
346 | domain_adaptation_images/webcam/images/mobile_phone/frame_0008.jpg 14
347 | domain_adaptation_images/webcam/images/mobile_phone/frame_0009.jpg 14
348 | domain_adaptation_images/webcam/images/mobile_phone/frame_0010.jpg 14
349 | domain_adaptation_images/webcam/images/mobile_phone/frame_0011.jpg 14
350 | domain_adaptation_images/webcam/images/mobile_phone/frame_0012.jpg 14
351 | domain_adaptation_images/webcam/images/mobile_phone/frame_0013.jpg 14
352 | domain_adaptation_images/webcam/images/mobile_phone/frame_0014.jpg 14
353 | domain_adaptation_images/webcam/images/mobile_phone/frame_0015.jpg 14
354 | domain_adaptation_images/webcam/images/mobile_phone/frame_0016.jpg 14
355 | domain_adaptation_images/webcam/images/mobile_phone/frame_0017.jpg 14
356 | domain_adaptation_images/webcam/images/mobile_phone/frame_0018.jpg 14
357 | domain_adaptation_images/webcam/images/mobile_phone/frame_0019.jpg 14
358 | domain_adaptation_images/webcam/images/mobile_phone/frame_0020.jpg 14
359 | domain_adaptation_images/webcam/images/mobile_phone/frame_0021.jpg 14
360 | domain_adaptation_images/webcam/images/mobile_phone/frame_0022.jpg 14
361 | domain_adaptation_images/webcam/images/mobile_phone/frame_0023.jpg 14
362 | domain_adaptation_images/webcam/images/mobile_phone/frame_0024.jpg 14
363 | domain_adaptation_images/webcam/images/mobile_phone/frame_0025.jpg 14
364 | domain_adaptation_images/webcam/images/mobile_phone/frame_0026.jpg 14
365 | domain_adaptation_images/webcam/images/mobile_phone/frame_0027.jpg 14
366 | domain_adaptation_images/webcam/images/mobile_phone/frame_0028.jpg 14
367 | domain_adaptation_images/webcam/images/mobile_phone/frame_0029.jpg 14
368 | domain_adaptation_images/webcam/images/mobile_phone/frame_0030.jpg 14
369 | domain_adaptation_images/webcam/images/monitor/frame_0001.jpg 15
370 | domain_adaptation_images/webcam/images/monitor/frame_0002.jpg 15
371 | domain_adaptation_images/webcam/images/monitor/frame_0003.jpg 15
372 | domain_adaptation_images/webcam/images/monitor/frame_0004.jpg 15
373 | domain_adaptation_images/webcam/images/monitor/frame_0005.jpg 15
374 | domain_adaptation_images/webcam/images/monitor/frame_0006.jpg 15
375 | domain_adaptation_images/webcam/images/monitor/frame_0007.jpg 15
376 | domain_adaptation_images/webcam/images/monitor/frame_0008.jpg 15
377 | domain_adaptation_images/webcam/images/monitor/frame_0009.jpg 15
378 | domain_adaptation_images/webcam/images/monitor/frame_0010.jpg 15
379 | domain_adaptation_images/webcam/images/monitor/frame_0011.jpg 15
380 | domain_adaptation_images/webcam/images/monitor/frame_0012.jpg 15
381 | domain_adaptation_images/webcam/images/monitor/frame_0013.jpg 15
382 | domain_adaptation_images/webcam/images/monitor/frame_0014.jpg 15
383 | domain_adaptation_images/webcam/images/monitor/frame_0015.jpg 15
384 | domain_adaptation_images/webcam/images/monitor/frame_0016.jpg 15
385 | domain_adaptation_images/webcam/images/monitor/frame_0017.jpg 15
386 | domain_adaptation_images/webcam/images/monitor/frame_0018.jpg 15
387 | domain_adaptation_images/webcam/images/monitor/frame_0019.jpg 15
388 | domain_adaptation_images/webcam/images/monitor/frame_0020.jpg 15
389 | domain_adaptation_images/webcam/images/monitor/frame_0021.jpg 15
390 | domain_adaptation_images/webcam/images/monitor/frame_0022.jpg 15
391 | domain_adaptation_images/webcam/images/monitor/frame_0023.jpg 15
392 | domain_adaptation_images/webcam/images/monitor/frame_0024.jpg 15
393 | domain_adaptation_images/webcam/images/monitor/frame_0025.jpg 15
394 | domain_adaptation_images/webcam/images/monitor/frame_0026.jpg 15
395 | domain_adaptation_images/webcam/images/monitor/frame_0027.jpg 15
396 | domain_adaptation_images/webcam/images/monitor/frame_0028.jpg 15
397 | domain_adaptation_images/webcam/images/monitor/frame_0029.jpg 15
398 | domain_adaptation_images/webcam/images/monitor/frame_0030.jpg 15
399 | domain_adaptation_images/webcam/images/monitor/frame_0031.jpg 15
400 | domain_adaptation_images/webcam/images/monitor/frame_0032.jpg 15
401 | domain_adaptation_images/webcam/images/monitor/frame_0033.jpg 15
402 | domain_adaptation_images/webcam/images/monitor/frame_0034.jpg 15
403 | domain_adaptation_images/webcam/images/monitor/frame_0035.jpg 15
404 | domain_adaptation_images/webcam/images/monitor/frame_0036.jpg 15
405 | domain_adaptation_images/webcam/images/monitor/frame_0037.jpg 15
406 | domain_adaptation_images/webcam/images/monitor/frame_0038.jpg 15
407 | domain_adaptation_images/webcam/images/monitor/frame_0039.jpg 15
408 | domain_adaptation_images/webcam/images/monitor/frame_0040.jpg 15
409 | domain_adaptation_images/webcam/images/monitor/frame_0041.jpg 15
410 | domain_adaptation_images/webcam/images/monitor/frame_0042.jpg 15
411 | domain_adaptation_images/webcam/images/monitor/frame_0043.jpg 15
412 | domain_adaptation_images/webcam/images/mouse/frame_0001.jpg 16
413 | domain_adaptation_images/webcam/images/mouse/frame_0002.jpg 16
414 | domain_adaptation_images/webcam/images/mouse/frame_0003.jpg 16
415 | domain_adaptation_images/webcam/images/mouse/frame_0004.jpg 16
416 | domain_adaptation_images/webcam/images/mouse/frame_0005.jpg 16
417 | domain_adaptation_images/webcam/images/mouse/frame_0006.jpg 16
418 | domain_adaptation_images/webcam/images/mouse/frame_0007.jpg 16
419 | domain_adaptation_images/webcam/images/mouse/frame_0008.jpg 16
420 | domain_adaptation_images/webcam/images/mouse/frame_0009.jpg 16
421 | domain_adaptation_images/webcam/images/mouse/frame_0010.jpg 16
422 | domain_adaptation_images/webcam/images/mouse/frame_0011.jpg 16
423 | domain_adaptation_images/webcam/images/mouse/frame_0012.jpg 16
424 | domain_adaptation_images/webcam/images/mouse/frame_0013.jpg 16
425 | domain_adaptation_images/webcam/images/mouse/frame_0014.jpg 16
426 | domain_adaptation_images/webcam/images/mouse/frame_0015.jpg 16
427 | domain_adaptation_images/webcam/images/mouse/frame_0016.jpg 16
428 | domain_adaptation_images/webcam/images/mouse/frame_0017.jpg 16
429 | domain_adaptation_images/webcam/images/mouse/frame_0018.jpg 16
430 | domain_adaptation_images/webcam/images/mouse/frame_0019.jpg 16
431 | domain_adaptation_images/webcam/images/mouse/frame_0020.jpg 16
432 | domain_adaptation_images/webcam/images/mouse/frame_0021.jpg 16
433 | domain_adaptation_images/webcam/images/mouse/frame_0022.jpg 16
434 | domain_adaptation_images/webcam/images/mouse/frame_0023.jpg 16
435 | domain_adaptation_images/webcam/images/mouse/frame_0024.jpg 16
436 | domain_adaptation_images/webcam/images/mouse/frame_0025.jpg 16
437 | domain_adaptation_images/webcam/images/mouse/frame_0026.jpg 16
438 | domain_adaptation_images/webcam/images/mouse/frame_0027.jpg 16
439 | domain_adaptation_images/webcam/images/mouse/frame_0028.jpg 16
440 | domain_adaptation_images/webcam/images/mouse/frame_0029.jpg 16
441 | domain_adaptation_images/webcam/images/mouse/frame_0030.jpg 16
442 | domain_adaptation_images/webcam/images/mug/frame_0001.jpg 17
443 | domain_adaptation_images/webcam/images/mug/frame_0002.jpg 17
444 | domain_adaptation_images/webcam/images/mug/frame_0003.jpg 17
445 | domain_adaptation_images/webcam/images/mug/frame_0004.jpg 17
446 | domain_adaptation_images/webcam/images/mug/frame_0005.jpg 17
447 | domain_adaptation_images/webcam/images/mug/frame_0006.jpg 17
448 | domain_adaptation_images/webcam/images/mug/frame_0007.jpg 17
449 | domain_adaptation_images/webcam/images/mug/frame_0008.jpg 17
450 | domain_adaptation_images/webcam/images/mug/frame_0009.jpg 17
451 | domain_adaptation_images/webcam/images/mug/frame_0010.jpg 17
452 | domain_adaptation_images/webcam/images/mug/frame_0011.jpg 17
453 | domain_adaptation_images/webcam/images/mug/frame_0012.jpg 17
454 | domain_adaptation_images/webcam/images/mug/frame_0013.jpg 17
455 | domain_adaptation_images/webcam/images/mug/frame_0014.jpg 17
456 | domain_adaptation_images/webcam/images/mug/frame_0015.jpg 17
457 | domain_adaptation_images/webcam/images/mug/frame_0016.jpg 17
458 | domain_adaptation_images/webcam/images/mug/frame_0017.jpg 17
459 | domain_adaptation_images/webcam/images/mug/frame_0018.jpg 17
460 | domain_adaptation_images/webcam/images/mug/frame_0019.jpg 17
461 | domain_adaptation_images/webcam/images/mug/frame_0020.jpg 17
462 | domain_adaptation_images/webcam/images/mug/frame_0021.jpg 17
463 | domain_adaptation_images/webcam/images/mug/frame_0022.jpg 17
464 | domain_adaptation_images/webcam/images/mug/frame_0023.jpg 17
465 | domain_adaptation_images/webcam/images/mug/frame_0024.jpg 17
466 | domain_adaptation_images/webcam/images/mug/frame_0025.jpg 17
467 | domain_adaptation_images/webcam/images/mug/frame_0026.jpg 17
468 | domain_adaptation_images/webcam/images/mug/frame_0027.jpg 17
469 | domain_adaptation_images/webcam/images/paper_notebook/frame_0001.jpg 18
470 | domain_adaptation_images/webcam/images/paper_notebook/frame_0002.jpg 18
471 | domain_adaptation_images/webcam/images/paper_notebook/frame_0003.jpg 18
472 | domain_adaptation_images/webcam/images/paper_notebook/frame_0004.jpg 18
473 | domain_adaptation_images/webcam/images/paper_notebook/frame_0005.jpg 18
474 | domain_adaptation_images/webcam/images/paper_notebook/frame_0006.jpg 18
475 | domain_adaptation_images/webcam/images/paper_notebook/frame_0007.jpg 18
476 | domain_adaptation_images/webcam/images/paper_notebook/frame_0008.jpg 18
477 | domain_adaptation_images/webcam/images/paper_notebook/frame_0009.jpg 18
478 | domain_adaptation_images/webcam/images/paper_notebook/frame_0010.jpg 18
479 | domain_adaptation_images/webcam/images/paper_notebook/frame_0011.jpg 18
480 | domain_adaptation_images/webcam/images/paper_notebook/frame_0012.jpg 18
481 | domain_adaptation_images/webcam/images/paper_notebook/frame_0013.jpg 18
482 | domain_adaptation_images/webcam/images/paper_notebook/frame_0014.jpg 18
483 | domain_adaptation_images/webcam/images/paper_notebook/frame_0015.jpg 18
484 | domain_adaptation_images/webcam/images/paper_notebook/frame_0016.jpg 18
485 | domain_adaptation_images/webcam/images/paper_notebook/frame_0017.jpg 18
486 | domain_adaptation_images/webcam/images/paper_notebook/frame_0018.jpg 18
487 | domain_adaptation_images/webcam/images/paper_notebook/frame_0019.jpg 18
488 | domain_adaptation_images/webcam/images/paper_notebook/frame_0020.jpg 18
489 | domain_adaptation_images/webcam/images/paper_notebook/frame_0021.jpg 18
490 | domain_adaptation_images/webcam/images/paper_notebook/frame_0022.jpg 18
491 | domain_adaptation_images/webcam/images/paper_notebook/frame_0023.jpg 18
492 | domain_adaptation_images/webcam/images/paper_notebook/frame_0024.jpg 18
493 | domain_adaptation_images/webcam/images/paper_notebook/frame_0025.jpg 18
494 | domain_adaptation_images/webcam/images/paper_notebook/frame_0026.jpg 18
495 | domain_adaptation_images/webcam/images/paper_notebook/frame_0027.jpg 18
496 | domain_adaptation_images/webcam/images/paper_notebook/frame_0028.jpg 18
497 | domain_adaptation_images/webcam/images/pen/frame_0001.jpg 19
498 | domain_adaptation_images/webcam/images/pen/frame_0002.jpg 19
499 | domain_adaptation_images/webcam/images/pen/frame_0003.jpg 19
500 | domain_adaptation_images/webcam/images/pen/frame_0004.jpg 19
501 | domain_adaptation_images/webcam/images/pen/frame_0005.jpg 19
502 | domain_adaptation_images/webcam/images/pen/frame_0006.jpg 19
503 | domain_adaptation_images/webcam/images/pen/frame_0007.jpg 19
504 | domain_adaptation_images/webcam/images/pen/frame_0008.jpg 19
505 | domain_adaptation_images/webcam/images/pen/frame_0009.jpg 19
506 | domain_adaptation_images/webcam/images/pen/frame_0010.jpg 19
507 | domain_adaptation_images/webcam/images/pen/frame_0011.jpg 19
508 | domain_adaptation_images/webcam/images/pen/frame_0012.jpg 19
509 | domain_adaptation_images/webcam/images/pen/frame_0013.jpg 19
510 | domain_adaptation_images/webcam/images/pen/frame_0014.jpg 19
511 | domain_adaptation_images/webcam/images/pen/frame_0015.jpg 19
512 | domain_adaptation_images/webcam/images/pen/frame_0016.jpg 19
513 | domain_adaptation_images/webcam/images/pen/frame_0017.jpg 19
514 | domain_adaptation_images/webcam/images/pen/frame_0018.jpg 19
515 | domain_adaptation_images/webcam/images/pen/frame_0019.jpg 19
516 | domain_adaptation_images/webcam/images/pen/frame_0020.jpg 19
517 | domain_adaptation_images/webcam/images/pen/frame_0021.jpg 19
518 | domain_adaptation_images/webcam/images/pen/frame_0022.jpg 19
519 | domain_adaptation_images/webcam/images/pen/frame_0023.jpg 19
520 | domain_adaptation_images/webcam/images/pen/frame_0024.jpg 19
521 | domain_adaptation_images/webcam/images/pen/frame_0025.jpg 19
522 | domain_adaptation_images/webcam/images/pen/frame_0026.jpg 19
523 | domain_adaptation_images/webcam/images/pen/frame_0027.jpg 19
524 | domain_adaptation_images/webcam/images/pen/frame_0028.jpg 19
525 | domain_adaptation_images/webcam/images/pen/frame_0029.jpg 19
526 | domain_adaptation_images/webcam/images/pen/frame_0030.jpg 19
527 | domain_adaptation_images/webcam/images/pen/frame_0031.jpg 19
528 | domain_adaptation_images/webcam/images/pen/frame_0032.jpg 19
529 | domain_adaptation_images/webcam/images/phone/frame_0001.jpg 20
530 | domain_adaptation_images/webcam/images/phone/frame_0002.jpg 20
531 | domain_adaptation_images/webcam/images/phone/frame_0003.jpg 20
532 | domain_adaptation_images/webcam/images/phone/frame_0004.jpg 20
533 | domain_adaptation_images/webcam/images/phone/frame_0005.jpg 20
534 | domain_adaptation_images/webcam/images/phone/frame_0006.jpg 20
535 | domain_adaptation_images/webcam/images/phone/frame_0007.jpg 20
536 | domain_adaptation_images/webcam/images/phone/frame_0008.jpg 20
537 | domain_adaptation_images/webcam/images/phone/frame_0009.jpg 20
538 | domain_adaptation_images/webcam/images/phone/frame_0010.jpg 20
539 | domain_adaptation_images/webcam/images/phone/frame_0011.jpg 20
540 | domain_adaptation_images/webcam/images/phone/frame_0012.jpg 20
541 | domain_adaptation_images/webcam/images/phone/frame_0013.jpg 20
542 | domain_adaptation_images/webcam/images/phone/frame_0014.jpg 20
543 | domain_adaptation_images/webcam/images/phone/frame_0015.jpg 20
544 | domain_adaptation_images/webcam/images/phone/frame_0016.jpg 20
545 | domain_adaptation_images/webcam/images/printer/frame_0001.jpg 21
546 | domain_adaptation_images/webcam/images/printer/frame_0002.jpg 21
547 | domain_adaptation_images/webcam/images/printer/frame_0003.jpg 21
548 | domain_adaptation_images/webcam/images/printer/frame_0004.jpg 21
549 | domain_adaptation_images/webcam/images/printer/frame_0005.jpg 21
550 | domain_adaptation_images/webcam/images/printer/frame_0006.jpg 21
551 | domain_adaptation_images/webcam/images/printer/frame_0007.jpg 21
552 | domain_adaptation_images/webcam/images/printer/frame_0008.jpg 21
553 | domain_adaptation_images/webcam/images/printer/frame_0009.jpg 21
554 | domain_adaptation_images/webcam/images/printer/frame_0010.jpg 21
555 | domain_adaptation_images/webcam/images/printer/frame_0011.jpg 21
556 | domain_adaptation_images/webcam/images/printer/frame_0012.jpg 21
557 | domain_adaptation_images/webcam/images/printer/frame_0013.jpg 21
558 | domain_adaptation_images/webcam/images/printer/frame_0014.jpg 21
559 | domain_adaptation_images/webcam/images/printer/frame_0015.jpg 21
560 | domain_adaptation_images/webcam/images/printer/frame_0016.jpg 21
561 | domain_adaptation_images/webcam/images/printer/frame_0017.jpg 21
562 | domain_adaptation_images/webcam/images/printer/frame_0018.jpg 21
563 | domain_adaptation_images/webcam/images/printer/frame_0019.jpg 21
564 | domain_adaptation_images/webcam/images/printer/frame_0020.jpg 21
565 | domain_adaptation_images/webcam/images/projector/frame_0001.jpg 22
566 | domain_adaptation_images/webcam/images/projector/frame_0002.jpg 22
567 | domain_adaptation_images/webcam/images/projector/frame_0003.jpg 22
568 | domain_adaptation_images/webcam/images/projector/frame_0004.jpg 22
569 | domain_adaptation_images/webcam/images/projector/frame_0005.jpg 22
570 | domain_adaptation_images/webcam/images/projector/frame_0006.jpg 22
571 | domain_adaptation_images/webcam/images/projector/frame_0007.jpg 22
572 | domain_adaptation_images/webcam/images/projector/frame_0008.jpg 22
573 | domain_adaptation_images/webcam/images/projector/frame_0009.jpg 22
574 | domain_adaptation_images/webcam/images/projector/frame_0010.jpg 22
575 | domain_adaptation_images/webcam/images/projector/frame_0011.jpg 22
576 | domain_adaptation_images/webcam/images/projector/frame_0012.jpg 22
577 | domain_adaptation_images/webcam/images/projector/frame_0013.jpg 22
578 | domain_adaptation_images/webcam/images/projector/frame_0014.jpg 22
579 | domain_adaptation_images/webcam/images/projector/frame_0015.jpg 22
580 | domain_adaptation_images/webcam/images/projector/frame_0016.jpg 22
581 | domain_adaptation_images/webcam/images/projector/frame_0017.jpg 22
582 | domain_adaptation_images/webcam/images/projector/frame_0018.jpg 22
583 | domain_adaptation_images/webcam/images/projector/frame_0019.jpg 22
584 | domain_adaptation_images/webcam/images/projector/frame_0020.jpg 22
585 | domain_adaptation_images/webcam/images/projector/frame_0021.jpg 22
586 | domain_adaptation_images/webcam/images/projector/frame_0022.jpg 22
587 | domain_adaptation_images/webcam/images/projector/frame_0023.jpg 22
588 | domain_adaptation_images/webcam/images/projector/frame_0024.jpg 22
589 | domain_adaptation_images/webcam/images/projector/frame_0025.jpg 22
590 | domain_adaptation_images/webcam/images/projector/frame_0026.jpg 22
591 | domain_adaptation_images/webcam/images/projector/frame_0027.jpg 22
592 | domain_adaptation_images/webcam/images/projector/frame_0028.jpg 22
593 | domain_adaptation_images/webcam/images/projector/frame_0029.jpg 22
594 | domain_adaptation_images/webcam/images/projector/frame_0030.jpg 22
595 | domain_adaptation_images/webcam/images/punchers/frame_0001.jpg 23
596 | domain_adaptation_images/webcam/images/punchers/frame_0002.jpg 23
597 | domain_adaptation_images/webcam/images/punchers/frame_0003.jpg 23
598 | domain_adaptation_images/webcam/images/punchers/frame_0004.jpg 23
599 | domain_adaptation_images/webcam/images/punchers/frame_0005.jpg 23
600 | domain_adaptation_images/webcam/images/punchers/frame_0006.jpg 23
601 | domain_adaptation_images/webcam/images/punchers/frame_0007.jpg 23
602 | domain_adaptation_images/webcam/images/punchers/frame_0008.jpg 23
603 | domain_adaptation_images/webcam/images/punchers/frame_0009.jpg 23
604 | domain_adaptation_images/webcam/images/punchers/frame_0010.jpg 23
605 | domain_adaptation_images/webcam/images/punchers/frame_0011.jpg 23
606 | domain_adaptation_images/webcam/images/punchers/frame_0012.jpg 23
607 | domain_adaptation_images/webcam/images/punchers/frame_0013.jpg 23
608 | domain_adaptation_images/webcam/images/punchers/frame_0014.jpg 23
609 | domain_adaptation_images/webcam/images/punchers/frame_0015.jpg 23
610 | domain_adaptation_images/webcam/images/punchers/frame_0016.jpg 23
611 | domain_adaptation_images/webcam/images/punchers/frame_0017.jpg 23
612 | domain_adaptation_images/webcam/images/punchers/frame_0018.jpg 23
613 | domain_adaptation_images/webcam/images/punchers/frame_0019.jpg 23
614 | domain_adaptation_images/webcam/images/punchers/frame_0020.jpg 23
615 | domain_adaptation_images/webcam/images/punchers/frame_0021.jpg 23
616 | domain_adaptation_images/webcam/images/punchers/frame_0022.jpg 23
617 | domain_adaptation_images/webcam/images/punchers/frame_0023.jpg 23
618 | domain_adaptation_images/webcam/images/punchers/frame_0024.jpg 23
619 | domain_adaptation_images/webcam/images/punchers/frame_0025.jpg 23
620 | domain_adaptation_images/webcam/images/punchers/frame_0026.jpg 23
621 | domain_adaptation_images/webcam/images/punchers/frame_0027.jpg 23
622 | domain_adaptation_images/webcam/images/ring_binder/frame_0001.jpg 24
623 | domain_adaptation_images/webcam/images/ring_binder/frame_0002.jpg 24
624 | domain_adaptation_images/webcam/images/ring_binder/frame_0003.jpg 24
625 | domain_adaptation_images/webcam/images/ring_binder/frame_0004.jpg 24
626 | domain_adaptation_images/webcam/images/ring_binder/frame_0005.jpg 24
627 | domain_adaptation_images/webcam/images/ring_binder/frame_0006.jpg 24
628 | domain_adaptation_images/webcam/images/ring_binder/frame_0007.jpg 24
629 | domain_adaptation_images/webcam/images/ring_binder/frame_0008.jpg 24
630 | domain_adaptation_images/webcam/images/ring_binder/frame_0009.jpg 24
631 | domain_adaptation_images/webcam/images/ring_binder/frame_0010.jpg 24
632 | domain_adaptation_images/webcam/images/ring_binder/frame_0011.jpg 24
633 | domain_adaptation_images/webcam/images/ring_binder/frame_0012.jpg 24
634 | domain_adaptation_images/webcam/images/ring_binder/frame_0013.jpg 24
635 | domain_adaptation_images/webcam/images/ring_binder/frame_0014.jpg 24
636 | domain_adaptation_images/webcam/images/ring_binder/frame_0015.jpg 24
637 | domain_adaptation_images/webcam/images/ring_binder/frame_0016.jpg 24
638 | domain_adaptation_images/webcam/images/ring_binder/frame_0017.jpg 24
639 | domain_adaptation_images/webcam/images/ring_binder/frame_0018.jpg 24
640 | domain_adaptation_images/webcam/images/ring_binder/frame_0019.jpg 24
641 | domain_adaptation_images/webcam/images/ring_binder/frame_0020.jpg 24
642 | domain_adaptation_images/webcam/images/ring_binder/frame_0021.jpg 24
643 | domain_adaptation_images/webcam/images/ring_binder/frame_0022.jpg 24
644 | domain_adaptation_images/webcam/images/ring_binder/frame_0023.jpg 24
645 | domain_adaptation_images/webcam/images/ring_binder/frame_0024.jpg 24
646 | domain_adaptation_images/webcam/images/ring_binder/frame_0025.jpg 24
647 | domain_adaptation_images/webcam/images/ring_binder/frame_0026.jpg 24
648 | domain_adaptation_images/webcam/images/ring_binder/frame_0027.jpg 24
649 | domain_adaptation_images/webcam/images/ring_binder/frame_0028.jpg 24
650 | domain_adaptation_images/webcam/images/ring_binder/frame_0029.jpg 24
651 | domain_adaptation_images/webcam/images/ring_binder/frame_0030.jpg 24
652 | domain_adaptation_images/webcam/images/ring_binder/frame_0031.jpg 24
653 | domain_adaptation_images/webcam/images/ring_binder/frame_0032.jpg 24
654 | domain_adaptation_images/webcam/images/ring_binder/frame_0033.jpg 24
655 | domain_adaptation_images/webcam/images/ring_binder/frame_0034.jpg 24
656 | domain_adaptation_images/webcam/images/ring_binder/frame_0035.jpg 24
657 | domain_adaptation_images/webcam/images/ring_binder/frame_0036.jpg 24
658 | domain_adaptation_images/webcam/images/ring_binder/frame_0037.jpg 24
659 | domain_adaptation_images/webcam/images/ring_binder/frame_0038.jpg 24
660 | domain_adaptation_images/webcam/images/ring_binder/frame_0039.jpg 24
661 | domain_adaptation_images/webcam/images/ring_binder/frame_0040.jpg 24
662 | domain_adaptation_images/webcam/images/ruler/frame_0001.jpg 25
663 | domain_adaptation_images/webcam/images/ruler/frame_0002.jpg 25
664 | domain_adaptation_images/webcam/images/ruler/frame_0003.jpg 25
665 | domain_adaptation_images/webcam/images/ruler/frame_0004.jpg 25
666 | domain_adaptation_images/webcam/images/ruler/frame_0005.jpg 25
667 | domain_adaptation_images/webcam/images/ruler/frame_0006.jpg 25
668 | domain_adaptation_images/webcam/images/ruler/frame_0007.jpg 25
669 | domain_adaptation_images/webcam/images/ruler/frame_0008.jpg 25
670 | domain_adaptation_images/webcam/images/ruler/frame_0009.jpg 25
671 | domain_adaptation_images/webcam/images/ruler/frame_0010.jpg 25
672 | domain_adaptation_images/webcam/images/ruler/frame_0011.jpg 25
673 | domain_adaptation_images/webcam/images/scissors/frame_0001.jpg 26
674 | domain_adaptation_images/webcam/images/scissors/frame_0002.jpg 26
675 | domain_adaptation_images/webcam/images/scissors/frame_0003.jpg 26
676 | domain_adaptation_images/webcam/images/scissors/frame_0004.jpg 26
677 | domain_adaptation_images/webcam/images/scissors/frame_0005.jpg 26
678 | domain_adaptation_images/webcam/images/scissors/frame_0006.jpg 26
679 | domain_adaptation_images/webcam/images/scissors/frame_0007.jpg 26
680 | domain_adaptation_images/webcam/images/scissors/frame_0008.jpg 26
681 | domain_adaptation_images/webcam/images/scissors/frame_0009.jpg 26
682 | domain_adaptation_images/webcam/images/scissors/frame_0010.jpg 26
683 | domain_adaptation_images/webcam/images/scissors/frame_0011.jpg 26
684 | domain_adaptation_images/webcam/images/scissors/frame_0012.jpg 26
685 | domain_adaptation_images/webcam/images/scissors/frame_0013.jpg 26
686 | domain_adaptation_images/webcam/images/scissors/frame_0014.jpg 26
687 | domain_adaptation_images/webcam/images/scissors/frame_0015.jpg 26
688 | domain_adaptation_images/webcam/images/scissors/frame_0016.jpg 26
689 | domain_adaptation_images/webcam/images/scissors/frame_0017.jpg 26
690 | domain_adaptation_images/webcam/images/scissors/frame_0018.jpg 26
691 | domain_adaptation_images/webcam/images/scissors/frame_0019.jpg 26
692 | domain_adaptation_images/webcam/images/scissors/frame_0020.jpg 26
693 | domain_adaptation_images/webcam/images/scissors/frame_0021.jpg 26
694 | domain_adaptation_images/webcam/images/scissors/frame_0022.jpg 26
695 | domain_adaptation_images/webcam/images/scissors/frame_0023.jpg 26
696 | domain_adaptation_images/webcam/images/scissors/frame_0024.jpg 26
697 | domain_adaptation_images/webcam/images/scissors/frame_0025.jpg 26
698 | domain_adaptation_images/webcam/images/speaker/frame_0001.jpg 27
699 | domain_adaptation_images/webcam/images/speaker/frame_0002.jpg 27
700 | domain_adaptation_images/webcam/images/speaker/frame_0003.jpg 27
701 | domain_adaptation_images/webcam/images/speaker/frame_0004.jpg 27
702 | domain_adaptation_images/webcam/images/speaker/frame_0005.jpg 27
703 | domain_adaptation_images/webcam/images/speaker/frame_0006.jpg 27
704 | domain_adaptation_images/webcam/images/speaker/frame_0007.jpg 27
705 | domain_adaptation_images/webcam/images/speaker/frame_0008.jpg 27
706 | domain_adaptation_images/webcam/images/speaker/frame_0009.jpg 27
707 | domain_adaptation_images/webcam/images/speaker/frame_0010.jpg 27
708 | domain_adaptation_images/webcam/images/speaker/frame_0011.jpg 27
709 | domain_adaptation_images/webcam/images/speaker/frame_0012.jpg 27
710 | domain_adaptation_images/webcam/images/speaker/frame_0013.jpg 27
711 | domain_adaptation_images/webcam/images/speaker/frame_0014.jpg 27
712 | domain_adaptation_images/webcam/images/speaker/frame_0015.jpg 27
713 | domain_adaptation_images/webcam/images/speaker/frame_0016.jpg 27
714 | domain_adaptation_images/webcam/images/speaker/frame_0017.jpg 27
715 | domain_adaptation_images/webcam/images/speaker/frame_0018.jpg 27
716 | domain_adaptation_images/webcam/images/speaker/frame_0019.jpg 27
717 | domain_adaptation_images/webcam/images/speaker/frame_0020.jpg 27
718 | domain_adaptation_images/webcam/images/speaker/frame_0021.jpg 27
719 | domain_adaptation_images/webcam/images/speaker/frame_0022.jpg 27
720 | domain_adaptation_images/webcam/images/speaker/frame_0023.jpg 27
721 | domain_adaptation_images/webcam/images/speaker/frame_0024.jpg 27
722 | domain_adaptation_images/webcam/images/speaker/frame_0025.jpg 27
723 | domain_adaptation_images/webcam/images/speaker/frame_0026.jpg 27
724 | domain_adaptation_images/webcam/images/speaker/frame_0027.jpg 27
725 | domain_adaptation_images/webcam/images/speaker/frame_0028.jpg 27
726 | domain_adaptation_images/webcam/images/speaker/frame_0029.jpg 27
727 | domain_adaptation_images/webcam/images/speaker/frame_0030.jpg 27
728 | domain_adaptation_images/webcam/images/stapler/frame_0001.jpg 28
729 | domain_adaptation_images/webcam/images/stapler/frame_0002.jpg 28
730 | domain_adaptation_images/webcam/images/stapler/frame_0003.jpg 28
731 | domain_adaptation_images/webcam/images/stapler/frame_0004.jpg 28
732 | domain_adaptation_images/webcam/images/stapler/frame_0005.jpg 28
733 | domain_adaptation_images/webcam/images/stapler/frame_0006.jpg 28
734 | domain_adaptation_images/webcam/images/stapler/frame_0007.jpg 28
735 | domain_adaptation_images/webcam/images/stapler/frame_0008.jpg 28
736 | domain_adaptation_images/webcam/images/stapler/frame_0009.jpg 28
737 | domain_adaptation_images/webcam/images/stapler/frame_0010.jpg 28
738 | domain_adaptation_images/webcam/images/stapler/frame_0011.jpg 28
739 | domain_adaptation_images/webcam/images/stapler/frame_0012.jpg 28
740 | domain_adaptation_images/webcam/images/stapler/frame_0013.jpg 28
741 | domain_adaptation_images/webcam/images/stapler/frame_0014.jpg 28
742 | domain_adaptation_images/webcam/images/stapler/frame_0015.jpg 28
743 | domain_adaptation_images/webcam/images/stapler/frame_0016.jpg 28
744 | domain_adaptation_images/webcam/images/stapler/frame_0017.jpg 28
745 | domain_adaptation_images/webcam/images/stapler/frame_0018.jpg 28
746 | domain_adaptation_images/webcam/images/stapler/frame_0019.jpg 28
747 | domain_adaptation_images/webcam/images/stapler/frame_0020.jpg 28
748 | domain_adaptation_images/webcam/images/stapler/frame_0021.jpg 28
749 | domain_adaptation_images/webcam/images/stapler/frame_0022.jpg 28
750 | domain_adaptation_images/webcam/images/stapler/frame_0023.jpg 28
751 | domain_adaptation_images/webcam/images/stapler/frame_0024.jpg 28
752 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0001.jpg 29
753 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0002.jpg 29
754 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0003.jpg 29
755 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0004.jpg 29
756 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0005.jpg 29
757 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0006.jpg 29
758 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0007.jpg 29
759 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0008.jpg 29
760 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0009.jpg 29
761 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0010.jpg 29
762 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0011.jpg 29
763 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0012.jpg 29
764 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0013.jpg 29
765 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0014.jpg 29
766 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0015.jpg 29
767 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0016.jpg 29
768 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0017.jpg 29
769 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0018.jpg 29
770 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0019.jpg 29
771 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0020.jpg 29
772 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0021.jpg 29
773 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0022.jpg 29
774 | domain_adaptation_images/webcam/images/tape_dispenser/frame_0023.jpg 29
775 | domain_adaptation_images/webcam/images/trash_can/frame_0001.jpg 30
776 | domain_adaptation_images/webcam/images/trash_can/frame_0002.jpg 30
777 | domain_adaptation_images/webcam/images/trash_can/frame_0003.jpg 30
778 | domain_adaptation_images/webcam/images/trash_can/frame_0004.jpg 30
779 | domain_adaptation_images/webcam/images/trash_can/frame_0005.jpg 30
780 | domain_adaptation_images/webcam/images/trash_can/frame_0006.jpg 30
781 | domain_adaptation_images/webcam/images/trash_can/frame_0007.jpg 30
782 | domain_adaptation_images/webcam/images/trash_can/frame_0008.jpg 30
783 | domain_adaptation_images/webcam/images/trash_can/frame_0009.jpg 30
784 | domain_adaptation_images/webcam/images/trash_can/frame_0010.jpg 30
785 | domain_adaptation_images/webcam/images/trash_can/frame_0011.jpg 30
786 | domain_adaptation_images/webcam/images/trash_can/frame_0012.jpg 30
787 | domain_adaptation_images/webcam/images/trash_can/frame_0013.jpg 30
788 | domain_adaptation_images/webcam/images/trash_can/frame_0014.jpg 30
789 | domain_adaptation_images/webcam/images/trash_can/frame_0015.jpg 30
790 | domain_adaptation_images/webcam/images/trash_can/frame_0016.jpg 30
791 | domain_adaptation_images/webcam/images/trash_can/frame_0017.jpg 30
792 | domain_adaptation_images/webcam/images/trash_can/frame_0018.jpg 30
793 | domain_adaptation_images/webcam/images/trash_can/frame_0019.jpg 30
794 | domain_adaptation_images/webcam/images/trash_can/frame_0020.jpg 30
795 | domain_adaptation_images/webcam/images/trash_can/frame_0021.jpg 30
796 | 


--------------------------------------------------------------------------------